def addHistogramControls(self, parent=None):
# set default settings
self.spinLowerThreshold = 0
self.spinLowerChecked = False
self.spinUpperThreshold = 0
self.spinUpperChecked = False
self.netOption = 0
self.dstWeight = 0
self.kNN = 0
self.andor = 0
self.matrix = None
self.excludeLimit = 2
self.percentil = 0
self.graph = None
self.graph_matrix = None
if parent is None:
parent = self.controlArea
boxGeneral = OWGUI.widgetBox(parent, box="Distance boundaries")
ribg = OWGUI.widgetBox(boxGeneral, None, orientation="horizontal", addSpace=False)
OWGUI.lineEdit(ribg, self, "spinLowerThreshold", "Lower", orientation='horizontal', callback=self.changeLowerSpin, valueType=float, enterPlaceholder=True, controlWidth=100)
OWGUI.lineEdit(ribg, self, "spinUpperThreshold", "Upper ", orientation='horizontal', callback=self.changeUpperSpin, valueType=float, enterPlaceholder=True, controlWidth=100)
ribg.layout().addStretch(1)
#ribg = OWGUI.radioButtonsInBox(boxGeneral, self, "andor", [], orientation='horizontal', callback = self.generateGraph)
#OWGUI.appendRadioButton(ribg, self, "andor", "OR", callback = self.generateGraph)
#b = OWGUI.appendRadioButton(ribg, self, "andor", "AND", callback = self.generateGraph)
#b.setEnabled(False)
#ribg.hide(False)
ribg = OWGUI.widgetBox(boxGeneral, None, orientation="horizontal", addSpace=False)
OWGUI.spin(ribg, self, "kNN", 0, 1000, 1, label="kNN ", orientation='horizontal', callback=self.generateGraph, callbackOnReturn=1, controlWidth=100)
OWGUI.doubleSpin(ribg, self, "percentil", 0, 100, 0.1, label="Percentile", orientation='horizontal', callback=self.setPercentil, callbackOnReturn=1, controlWidth=100)
ribg.layout().addStretch(1)
# Options
self.attrColor = ""
ribg = OWGUI.radioButtonsInBox(parent, self, "netOption", [], "Options", callback=self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "All vertices", callback=self.generateGraph)
hb = OWGUI.widgetBox(ribg, None, orientation="horizontal", addSpace=False)
OWGUI.appendRadioButton(ribg, self, "netOption", "Large components only. Min nodes:", insertInto=hb, callback=self.generateGraph)
OWGUI.spin(hb, self, "excludeLimit", 2, 100, 1, callback=(lambda h=True: self.generateGraph(h)))
OWGUI.appendRadioButton(ribg, self, "netOption", "Largest connected component only", callback=self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "Connected component with vertex")
self.attribute = None
self.attributeCombo = OWGUI.comboBox(parent, self, "attribute", box="Filter attribute", orientation='horizontal')#, callback=self.setVertexColor)
ribg = OWGUI.radioButtonsInBox(parent, self, "dstWeight", [], "Distance -> Weight", callback=self.generateGraph)
hb = OWGUI.widgetBox(ribg, None, orientation="horizontal", addSpace=False)
OWGUI.appendRadioButton(ribg, self, "dstWeight", "Weight := distance", insertInto=hb, callback=self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "dstWeight", "Weight := 1 - distance", insertInto=hb, callback=self.generateGraph)
self.label = ''
self.searchString = OWGUI.lineEdit(self.attributeCombo.box, self, "label", callback=self.setSearchStringTimer, callbackOnType=True)
self.searchStringTimer = QTimer(self)
self.connect(self.searchStringTimer, SIGNAL("timeout()"), self.generateGraph)
if str(self.netOption) != '3':
self.attributeCombo.box.setEnabled(False)
def __init__(self, parent=None, signalManager=None,
title="Earth"):
OWWidget.__init__(self, parent, signalManager, title,
wantMainArea=False)
self.inputs = [("Data", Orange.data.Table, self.set_data),
("Preprocessor", PreprocessedLearner,
self.set_preprocessor)]
self.outputs = [("Learner", earth.EarthLearner, Default),
("Predictor", earth.EarthClassifier, Default),
("Basis Matrix", Orange.data.Table)]
self.name = "Earth Learner"
self.degree = 1
self.terms = 21
self.penalty = 2
self.loadSettings()
#####
# GUI
#####
OWGUI.lineEdit(self.controlArea, self, "name",
box="Learner/Classifier Name",
tooltip="Name for the learner/predictor")
box = OWGUI.widgetBox(self.controlArea, "Forward Pass", addSpace=True)
OWGUI.spin(box, self, "degree", 1, 3, step=1,
label="Max. term degree",
tooltip="Maximum degree of the terms derived "
"(number of hinge functions).")
s = OWGUI.spin(box, self, "terms", 1, 200, step=1,
label="Max. terms",
tooltip="Maximum number of terms derived in the "
"forward pass.")
s.control.setSpecialValueText("Automatic")
box = OWGUI.widgetBox(self.controlArea, "Pruning Pass", addSpace=True)
OWGUI.doubleSpin(box, self, "penalty", min=0.0, max=10.0, step=0.25,
label="Knot penalty")
OWGUI.button(self.controlArea, self, "&Apply",
callback=self.apply)
self.data = None
self.preprocessor = None
self.resize(300, 200)
self.apply()
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self, parent, signalManager, "Pubmed Network View", wantMainArea=0)
self.inputs = []
self.outputs = [("Nx View", Orange.network.NxView)]
self._nhops = 2
self._edge_threshold = 0.5
self._n_max_neighbors = 20
self.selected_titles = []
self.titles = []
self.filter = ""
self.ids = []
self._selected_nodes = []
self._algorithm = 0
self._k_algorithm = 0.3
self.loadSettings()
box = OWGUI.widgetBox(self.controlArea, "Paper Selection", orientation="vertical")
OWGUI.lineEdit(box, self, "filter", callback=self.filter_list, callbackOnType=True)
self.list_titles = OWGUI.listBox(
box, self, "selected_titles", "titles", selectionMode=QListWidget.MultiSelection, callback=self.update_view
)
OWGUI.separator(self.controlArea)
box_pref = OWGUI.widgetBox(self.controlArea, "Preferences", orientation="vertical")
OWGUI.spin(box_pref, self, "_nhops", 1, 6, 1, label="Number of hops: ", callback=self.update_view)
OWGUI.spin(
box_pref, self, "_n_max_neighbors", 1, 100, 1, label="Max number of neighbors: ", callback=self.update_view
)
OWGUI.doubleSpin(
box_pref, self, "_edge_threshold", 0, 1, step=0.01, label="Edge threshold: ", callback=self.update_view
)
OWGUI.separator(self.controlArea)
box_alg = OWGUI.widgetBox(self.controlArea, "Interest Propagation Algorithm", orientation="vertical")
radio_box = OWGUI.radioButtonsInBox(box_alg, self, "_algorithm", [], callback=self.update_view)
OWGUI.appendRadioButton(radio_box, self, "_algorithm", "Without Clustering", callback=self.update_view)
OWGUI.doubleSpin(
OWGUI.indentedBox(radio_box),
self,
"_k_algorithm",
0,
1,
step=0.01,
label="Parameter k: ",
callback=self.update_view,
)
OWGUI.appendRadioButton(radio_box, self, "_algorithm", "With Clustering", callback=self.update_view)
self.inside_view = PubmedNetworkView(self)
self.send("Nx View", self.inside_view)
def addHistogramControls(self, parent=None):
# set default settings
self.spinLowerThreshold = 0
self.spinLowerChecked = False
self.spinUpperThreshold = 0
self.spinUpperChecked = False
self.netOption = 0
self.dstWeight = 0
self.kNN = 0
self.andor = 0
self.matrix = None
self.excludeLimit = 1
self.percentil = 0
if parent is None:
parent = self.controlArea
boxGeneral = OWGUI.widgetBox(parent, box = "Distance boundaries")
OWGUI.lineEdit(boxGeneral, self, "spinLowerThreshold", "Lower:", orientation='horizontal', callback=self.changeLowerSpin, valueType=float)
OWGUI.lineEdit(boxGeneral, self, "spinUpperThreshold", "Upper:", orientation='horizontal', callback=self.changeUpperSpin, valueType=float)
ribg = OWGUI.radioButtonsInBox(boxGeneral, self, "andor", [], orientation='horizontal', callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "andor", "OR", callback = self.generateGraph)
b = OWGUI.appendRadioButton(ribg, self, "andor", "AND", callback = self.generateGraph)
b.setEnabled(False)
OWGUI.spin(boxGeneral, self, "kNN", 0, 1000, 1, label="kNN:", orientation='horizontal', callback=self.generateGraph)
OWGUI.doubleSpin(boxGeneral, self, "percentil", 0, 100, 0.1, label="Percentil:", orientation='horizontal', callback=self.setPercentil, callbackOnReturn=1)
# Options
self.attrColor = ""
ribg = OWGUI.radioButtonsInBox(parent, self, "netOption", [], "Options", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "All vertices", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "Exclude small components", callback = self.generateGraph)
OWGUI.spin(OWGUI.indentedBox(ribg), self, "excludeLimit", 1, 100, 1, label="Less vertices than: ", callback = (lambda h=True: self.generateGraph(h)))
OWGUI.appendRadioButton(ribg, self, "netOption", "Largest connected component only", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "Connected component with vertex")
self.attribute = None
self.attributeCombo = OWGUI.comboBox(ribg, self, "attribute", box = "Filter attribute")#, callback=self.setVertexColor)
ribg = OWGUI.radioButtonsInBox(parent, self, "dstWeight", [], "Distance -> Weight", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "dstWeight", "Weight := distance", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "dstWeight", "Weight := 1 - distance", callback = self.generateGraph)
self.label = ''
self.searchString = OWGUI.lineEdit(self.attributeCombo.box, self, "label", callback=self.setSearchStringTimer, callbackOnType=True)
self.searchStringTimer = QTimer(self)
self.connect(self.searchStringTimer, SIGNAL("timeout()"), self.generateGraph)
if str(self.netOption) != '3':
self.attributeCombo.box.setEnabled(False)
def addHistogramControls(self, parent=None):
# set default settings
self.spinLowerThreshold = 0
self.spinLowerChecked = False
self.spinUpperThreshold = 0
self.spinUpperChecked = False
self.netOption = 0
self.dstWeight = 0
self.kNN = 0
self.andor = 0
self.matrix = None
self.excludeLimit = 1
self.percentil = 0
if parent is None:
parent = self.controlArea
boxGeneral = OWGUI.widgetBox(parent, box = "Distance boundaries")
OWGUI.lineEdit(boxGeneral, self, "spinLowerThreshold", "Lower:", orientation='horizontal', callback=self.changeLowerSpin, valueType=float)
OWGUI.lineEdit(boxGeneral, self, "spinUpperThreshold", "Upper:", orientation='horizontal', callback=self.changeUpperSpin, valueType=float)
ribg = OWGUI.radioButtonsInBox(boxGeneral, self, "andor", [], orientation='horizontal', callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "andor", "OR", callback = self.generateGraph)
b = OWGUI.appendRadioButton(ribg, self, "andor", "AND", callback = self.generateGraph)
b.setEnabled(False)
OWGUI.spin(boxGeneral, self, "kNN", 0, 1000, 1, label="kNN:", orientation='horizontal', callback=self.generateGraph)
OWGUI.doubleSpin(boxGeneral, self, "percentil", 0, 100, 0.1, label="Percentil:", orientation='horizontal', callback=self.setPercentil, callbackOnReturn=1)
# Options
self.attrColor = ""
ribg = OWGUI.radioButtonsInBox(parent, self, "netOption", [], "Options", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "All vertices", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "Exclude small components", callback = self.generateGraph)
OWGUI.spin(OWGUI.indentedBox(ribg), self, "excludeLimit", 1, 100, 1, label="Less vertices than: ", callback = (lambda h=True: self.generateGraph(h)))
OWGUI.appendRadioButton(ribg, self, "netOption", "Largest connected component only", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "netOption", "Connected component with vertex")
self.attribute = None
self.attributeCombo = OWGUI.comboBox(ribg, self, "attribute", box = "Filter attribute")#, callback=self.setVertexColor)
ribg = OWGUI.radioButtonsInBox(parent, self, "dstWeight", [], "Distance -> Weight", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "dstWeight", "Weight := distance", callback = self.generateGraph)
OWGUI.appendRadioButton(ribg, self, "dstWeight", "Weight := 1 - distance", callback = self.generateGraph)
self.label = ''
self.searchString = OWGUI.lineEdit(self.attributeCombo.box, self, "label", callback=self.setSearchStringTimer, callbackOnType=True)
self.searchStringTimer = QTimer(self)
self.connect(self.searchStringTimer, SIGNAL("timeout()"), self.generateGraph)
if str(self.netOption) != '3':
self.attributeCombo.box.setEnabled(False)
def __init__(self, parent=None, signalManager=None, name="Fragmenter"):
OWWidget.__init__(self, parent, signalManager, name, wantMainArea=False)
self.inputs = [("Active chemicals", ExampleTable, self.setActive), ("Inactive chemicals", ExampleTable, self.setInactive)]
self.outputs = [("Fragments", ExampleTable)]
self.minSupport = 0.2
self.maxSupport = 0.2
self.activeSmilesAttr = 0
self.inactiveSmilesAttr = 0
self.loadSettings()
# GUI
box = OWGUI.widgetBox(self.controlArea, "Active Chemicals Set")
OWGUI.doubleSpin(box, self, "minSupport", 0.05, 0.95, step=0.05,
label="Min. active frequency",
tooltip="Minimal fragment frequency in the active chemicals set.",
callback=self.updateFreq)
self.activeSmilesAttrCB = OWGUI.comboBox(box, self, "activeSmilesAttr",
label="SMILES attribute",
callback=self.updateAttrs)
box = OWGUI.widgetBox(self.controlArea, "Inactive Chemicals Set")
OWGUI.doubleSpin(box, self, "maxSupport", 0.05, 0.95, step=0.05,
label="Max. inactive frequency",
tooltip="Maximal fragment frequency in the inactive chemicals set.",
callback=self.updateFreq)
self.inactiveSmilesAttrCB = OWGUI.comboBox(box, self, "inactiveSmilesAttr",
label="SMILES attribute",
callback=self.updateAttrs)
OWGUI.button(self.controlArea, self, "Run",
callback=self.run)
self.activeData = None
self.inactiveData = None
self.activeDataAttrs = []
self.inactiveDataAttrs = []
self.resize(100, 100)
def __init__(self, parent=None, signalManager = None):
OWWidget.__init__(self, parent, signalManager, "Predictions")
self.callbackDeposit = []
self.inputs = [("Examples", ExampleTable, self.setData), ("Predictors", orange.Classifier, self.setPredictor, Multiple)]
self.outputs = [("Predictions", ExampleTable)]
self.predictors = {}
# saveble settings
self.ShowAttributeMethod = 0
self.classes = []
self.selectedClasses = []
self.loadSettings()
self.datalabel = "N/A"
self.predictorlabel = "N/A"
self.tasklabel = "N/A"
self.outvar = None # current output variable (set by the first predictor/data set send in)
self.classifications = []
self.rindx = None
self.data = None
self.verbose = 0
self.nVarImportance = 0
# GUI - Options
# Options - classification
ibox = OWGUI.widgetBox(self.controlArea, "Info")
OWGUI.label(ibox, self, "Data: %(datalabel)s")
OWGUI.label(ibox, self, "Predictors: %(predictorlabel)s")
OWGUI.label(ibox, self, "Task: %(tasklabel)s")
OWGUI.separator(ibox)
OWGUI.label(ibox, self, "Predictions can be viewed with the 'Data Table'\nand saved with the 'Save' widget!")
OWGUI.separator(self.controlArea)
self.copt = OWGUI.widgetBox(self.controlArea,"Probabilities (classification)")
self.copt.setDisabled(1)
self.lbcls = OWGUI.listBox(self.copt, self, "selectedClasses", "classes",
selectionMode=QListWidget.MultiSelection)
self.lbcls.setFixedHeight(50)
OWGUI.separator(self.controlArea)
self.VarImportanceBox = OWGUI.doubleSpin(self.controlArea, self, "nVarImportance", 0,9999999,1, label="Variable importance", orientation="horizontal", tooltip="The number of variables to report for each prediction.")
OWGUI.checkBox(self.controlArea, self, 'verbose', 'Verbose', tooltip='Show detailed info while predicting. This will slow down the predictions process!')
OWGUI.separator(self.controlArea)
self.outbox = OWGUI.widgetBox(self.controlArea,"Output")
self.apply = OWGUI.button(self.outbox, self, "Apply", callback=self.sendpredictions)
OWGUI.rubber(self.controlArea)
self.adjustSize()
def __init__ (self, parent=None, signalManager = None, name = "Logistic regression"):
OWWidget.__init__(self, parent, signalManager, name, wantMainArea = 0, resizingEnabled = 0)
self.inputs = [("Data", ExampleTable, self.setData), ("Preprocess", PreprocessedLearner, self.setPreprocessor)]
self.outputs = [("Learner", orange.Learner), ("Classifier", orange.Classifier), ("Features", list)]
self.regularizations = [ Orange.classification.logreg.LibLinearLogRegLearner.L2R_LR, Orange.classification.logreg.LibLinearLogRegLearner.L1R_LR ]
self.regularizationsStr = [ "L2 (squared weights)", "L1 (absolute weights)" ]
self.name = "Logistic regression"
self.normalization = True
self.C = 1.
self.regularization = 0
self.data = None
self.preprocessor = None
self.loadSettings()
OWGUI.lineEdit(self.controlArea, self, 'name', box='Learner/Classifier Name', tooltip='Name to be used by other widgets to identify your learner/classifier.')
OWGUI.separator(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Regularization")
self.regularizationCombo = OWGUI.comboBox(box, self, "regularization", items=self.regularizationsStr)
cset = OWGUI.doubleSpin(box, self, "C", 0.01, 512.0, 0.1,
decimals=2,
addToLayout=True,
label="Training error cost (C)",
alignment=Qt.AlignRight,
tooltip= "Weight of log-loss term (higher C means better fit on the training data)."),
OWGUI.separator(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Preprocessing")
OWGUI.checkBox(box, self, "normalization",
label="Normalize data",
tooltip="Normalize data before learning.")
OWGUI.separator(self.controlArea)
applyButton = OWGUI.button(self.controlArea, self, "&Apply", callback=self.applyLearner, default=True)
OWGUI.rubber(self.controlArea)
#self.adjustSize()
self.applyLearner()
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self, parent, signalManager, 'Nx Clustering')
self.inputs = [("Network", Orange.network.Graph,
self.setNetwork, Default)]
self.outputs = [("Network", Orange.network.Graph)]
self.net = None
self.method = 0
self.iterationHistory = 0
self.autoApply = 0
self.iterations = 1000
self.hop_attenuation = 0.1
self.loadSettings()
OWGUI.spin(self.controlArea, self, "iterations", 1,
100000, 1, label="Iterations: ")
ribg = OWGUI.radioButtonsInBox(self.controlArea, self, "method",
[], "Method", callback=self.cluster)
OWGUI.appendRadioButton(ribg, self, "method",
"Label propagation clustering (Raghavan et al., 2007)",
callback=self.cluster)
OWGUI.appendRadioButton(ribg, self, "method",
"Label propagation clustering (Leung et al., 2009)",
callback=self.cluster)
OWGUI.doubleSpin(OWGUI.indentedBox(ribg), self, "hop_attenuation",
0, 1, 0.01, label="Hop attenuation (delta): ")
self.info = OWGUI.widgetLabel(self.controlArea, ' ')
OWGUI.checkBox(self.controlArea, self, "iterationHistory",
"Append clustering data on each iteration",
callback=self.cluster)
OWGUI.checkBox(self.controlArea, self, "autoApply",
"Commit automatically")
OWGUI.button(self.controlArea, self, "Commit",
callback=lambda b=True: self.cluster(b))
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self, parent, signalManager, 'Pubmed Network View', wantMainArea=0)
self.inputs = []
self.outputs = [("Nx View", Orange.network.NxView)]
self._nhops = 2
self._edge_threshold = 0.5
self._n_max_neighbors = 20
self.selected_titles = []
self.titles = []
self.filter = ''
self.ids = []
self._selected_nodes = []
self._algorithm = 0
self._k_algorithm = 0.3
self.loadSettings()
box = OWGUI.widgetBox(self.controlArea, "Paper Selection", orientation="vertical")
OWGUI.lineEdit(box, self, "filter", callback=self.filter_list, callbackOnType=True)
self.list_titles = OWGUI.listBox(box, self, "selected_titles", "titles", selectionMode=QListWidget.MultiSelection, callback=self.update_view)
OWGUI.separator(self.controlArea)
box_pref = OWGUI.widgetBox(self.controlArea, "Preferences", orientation="vertical")
OWGUI.spin(box_pref, self, "_nhops", 1, 6, 1, label="Number of hops: ", callback=self.update_view)
OWGUI.spin(box_pref, self, "_n_max_neighbors", 1, 100, 1, label="Max number of neighbors: ", callback=self.update_view)
OWGUI.doubleSpin(box_pref, self, "_edge_threshold", 0, 1, step=0.01, label="Edge threshold: ", callback=self.update_view)
OWGUI.separator(self.controlArea)
box_alg = OWGUI.widgetBox(self.controlArea, "Interest Propagation Algorithm", orientation="vertical")
radio_box = OWGUI.radioButtonsInBox(box_alg, self, "_algorithm", [], callback=self.update_view)
OWGUI.appendRadioButton(radio_box, self, "_algorithm", "Without Clustering", callback=self.update_view)
OWGUI.doubleSpin(OWGUI.indentedBox(radio_box), self, "_k_algorithm", 0, 1, step=0.01, label="Parameter k: ", callback=self.update_view)
OWGUI.appendRadioButton(radio_box, self, "_algorithm", "With Clustering", callback=self.update_view)
self.inside_view = PubmedNetworkView(self)
self.send("Nx View", self.inside_view)
def __init__(self, parent=None, signalManager=None, name="SVM"):
OWWidget.__init__(self, parent, signalManager, name)
self.inputs=[("Example Table", ExampleTable, self.setData)]
self.outputs=[("Learner", orange.Learner),("Classifier", orange.Classifier)] # ,("Support Vectors", ExampleTable)]
self.data = None
self.name="Support Vector Machine"
self.nMin = -1
self.nMax = 1
self.nClassMin = -1
self.nClassMax = 1
self.svm_typeGUI = None
self.stopCritGUI = None
self.priorsGUI = ""
#Read default parameters from AZOrangeConfig.py file
for par in ("kernel_type", "svm_type","gamma","C","p","epsC","epsR","stopCrit",\
"maxIter","nu","scaleData","scaleClass","priors","coef0","degree"):
setattr(self, par, AZOC.CVSVMDEFAULTDICT[par])
if self.svm_type in (103,104):
self.eps = self.epsR
else:
self.eps = self.epsC
self.setSVMTypeGUI(self.svm_type)
self.setStopCritGUI(self.stopCrit)
self.setPriorsGUI()#self.priors)
##scJS Added parameter needed for saving an SVM model
self.modelFile = os.path.join(os.getcwd(),"SVM.model")
##ecJS
OWGUI.lineEdit(self.controlArea, self, 'name', box='Learner/Classifier Name', \
tooltip='Name to be used by other widgets to identify your learner/classifier.<br>This should be a unique name!')
## scLC
# Do not use the one class svm (type 2).
self.svmTypeBox=b=OWGUI.widgetBox(self.controlArea,"SVM Type")
self.svmTyperadio = OWGUI.radioButtonsInBox(b, self, "svm_typeGUI", btnLabels=["C-SVC", "nu-SVC","epsilon-SVR", "nu-SVR"], callback=self.setSVMTypeLearner)
## ecLC
self.kernelBox=b=OWGUI.widgetBox(self.controlArea, "Kernel")
self.kernelradio = OWGUI.radioButtonsInBox(b, self, "kernel_type", btnLabels=["Linear, x.y", "Polynomial, (g*x.y+c)^d",
"RBF, exp(-g*(x-y).(x-y))", "Sigmoid, tanh(g*x.y+c)"], callback=self.changeKernel)
self.gcd = OWGUI.widgetBox(b, orientation="horizontal")
self.leg = OWGUI.doubleSpin(self.gcd, self, "gamma",0.0,10.0,0.0001, label="gamma: ", orientation="horizontal", callback=self.changeKernel)
self.led = OWGUI.doubleSpin(self.gcd, self, "coef0", 0.0,10.0,0.0001, label=" coef0: ", orientation="horizontal", callback=self.changeKernel)
self.lec = OWGUI.doubleSpin(self.gcd, self, "degree", 0.0,10.0,0.5, label=" degree: ", orientation="horizontal", callback=self.changeKernel)
OWGUI.separator(self.controlArea)
#self.layout=QVBoxLayout(self.mainArea)
self.optionsBox=b=OWGUI.widgetBox(self.mainArea, "Options", addSpace = True)
self.cBox = OWGUI.doubleSpin(b,self, "C", 0.0, 512.0, 0.5, label="Model complexity (C)", orientation="horizontal", tooltip="Penalty cost")
self.cBox.control.setDecimals(3)
self.pBox = OWGUI.doubleSpin(b,self, "p", 0.0, 10.0, 0.1, label="Tolerance (p)", orientation="horizontal", tooltip="zero-loss width zone")
self.pBox.control.setDecimals(3)
self.nuBox = OWGUI.doubleSpin(b, self, "nu", 0.0,1.0,0.1, label="Complexity bound (nu)", orientation="horizontal", tooltip="Upper bound on the ratio of support vectors")
self.nuBox.control.setDecimals(3)
self.priorBox = OWGUI.lineEdit(b, self, 'priorsGUI', box='Class weights', \
tooltip='Ex: POS:2 , NEG:1')
self.stopCritBox=OWGUI.widgetBox(self.optionsBox,"Stop Criteria")
self.stopCritradio = OWGUI.radioButtonsInBox(self.stopCritBox, self, "stopCritGUI", btnLabels=["Maximum number of iterations (maxIter)", "Numeric precision (eps)"], callback=self.setStopCritLearner)
OWGUI.doubleSpin(self.stopCritBox,self, "eps", 0.0, 0.5, 0.001, label="Numeric precision (eps)", orientation="horizontal")
OWGUI.doubleSpin(self.stopCritBox,self, "maxIter", 0, 100000, 1, label="Maximum number of iterations (maxIter)", orientation="horizontal")
self.scaleDataBox=OWGUI.checkBox(b, self, "scaleData", "Scale Data", tooltip="Scales the training data, and the input examples",callback = self.changeScaling)
self.scaleClassBox=OWGUI.checkBox(OWGUI.indentedBox(b), self, "scaleClass", "Scale Class variable", tooltip="Scales the class variable of training data, and the output predictions",callback = self.changeScaling)
#self.scalenMin = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nMin', 'Var min:', orientation="horizontal", tooltip='Variables scaling minimum')
#self.scalenMax = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nMax', 'Var max:', orientation="horizontal", tooltip='Variables scaling maximum')
#OWGUI.separator(OWGUI.indentedBox(b))
self.scalenClassMin = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nClassMin', 'Class min:', orientation="horizontal", tooltip='Class scaling minimum')
self.scalenClassMax = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nClassMax', 'Class max:', orientation="horizontal", tooltip='Class scaling maximum')
OWGUI.separator(self.controlArea)
#self.paramButton=OWGUI.button(self.controlArea, self, "Automatic parameter search", callback=self.parameterSearch,
# tooltip="Automaticaly searches for parameters that optimize classifier acuracy")
self.optionsBox.adjustSize()
#self.layout.add(self.optionsBox)
#OWGUI.separator(self.controlArea)
OWGUI.button(self.controlArea, self,"&Apply", callback=self.applySettings)
##scJS Adding saving of svm models
# Get desired location of model file
boxFile = OWGUI.widgetBox(self.controlArea, "Path for saving Model", addSpace = True, orientation=0)
L1 = OWGUI.lineEdit(boxFile, self, "modelFile", labelWidth=80, orientation = "horizontal", \
tooltip = "Once a model is created (connect this widget with a data widget), \nit can be saved by giving a \
file name here and clicking the save button.")
#L1.setMinimumWidth(200)
button = OWGUI.button(boxFile, self, '...', callback = self.browseFile, disabled=0,tooltip = "Choose the dir where to save. After chosen, add a name for the model file!")
#button.setMaximumWidth(25)
# Save model
OWGUI.button(self.controlArea, self,"&Save SVM model", callback=self.saveModel)
#self.adjustSize()
#self.loadSettings()
self.setStopCritLearner()
self.setSVMTypeLearner()
self.changeKernel()
self.applySettings()
self.resize(500,400)
def __init__(self, parent=None, signalManager=None, title="PCA"):
OWWidget.__init__(self, parent, signalManager, title, wantGraph=True)
self.inputs = [("Input Data", Orange.data.Table, self.set_data)]
self.outputs = [("Transformed Data", Orange.data.Table, Default),
("Eigen Vectors", Orange.data.Table)]
self.standardize = True
self.max_components = 0
self.variance_covered = 100.0
self.use_generalized_eigenvectors = False
self.auto_commit = False
self.loadSettings()
self.data = None
self.changed_flag = False
#####
# GUI
#####
grid = QGridLayout()
box = OWGUI.widgetBox(self.controlArea,
"Components Selection",
orientation=grid)
label1 = QLabel("Max components", box)
grid.addWidget(label1, 1, 0)
sb1 = OWGUI.spin(box,
self,
"max_components",
0,
1000,
tooltip="Maximum number of components",
callback=self.on_update,
addToLayout=False,
keyboardTracking=False)
self.max_components_spin = sb1.control
self.max_components_spin.setSpecialValueText("All")
grid.addWidget(sb1.control, 1, 1)
label2 = QLabel("Variance covered", box)
grid.addWidget(label2, 2, 0)
sb2 = OWGUI.doubleSpin(box,
self,
"variance_covered",
1.0,
100.0,
1.0,
tooltip="Percent of variance covered.",
callback=self.on_update,
decimals=1,
addToLayout=False,
keyboardTracking=False)
sb2.control.setSuffix("%")
grid.addWidget(sb2.control, 2, 1)
OWGUI.rubber(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Commit")
cb = OWGUI.checkBox(box, self, "auto_commit", "Commit on any change")
b = OWGUI.button(box, self, "Commit", callback=self.update_components)
OWGUI.setStopper(self, b, cb, "changed_flag", self.update_components)
self.plot = Graph()
canvas = self.plot.canvas()
canvas.setFrameStyle(QFrame.StyledPanel)
self.mainArea.layout().addWidget(self.plot)
self.plot.setAxisTitle(QwtPlot.yLeft, "Proportion of Variance")
self.plot.setAxisTitle(QwtPlot.xBottom, "Principal Components")
self.plot.setAxisScale(QwtPlot.yLeft, 0.0, 1.0)
self.plot.enableGridXB(True)
self.plot.enableGridYL(True)
self.plot.setGridColor(Qt.lightGray)
self.variance_curve = plot_curve("Variance",
pen=QPen(Qt.red, 2),
symbol=QwtSymbol.NoSymbol,
xaxis=QwtPlot.xBottom,
yaxis=QwtPlot.yLeft)
self.cumulative_variance_curve = plot_curve("Cumulative Variance",
pen=QPen(Qt.darkYellow, 2),
symbol=QwtSymbol.NoSymbol,
xaxis=QwtPlot.xBottom,
yaxis=QwtPlot.yLeft)
self.variance_curve.attach(self.plot)
self.cumulative_variance_curve.attach(self.plot)
self.selection_tool = CutoffControler(parent=self.plot.canvas())
self.selection_tool.cutoffMoved.connect(self.on_cutoff_moved)
self.graphButton.clicked.connect(self.saveToFile)
self.components = None
self.variances = None
self.variances_sum = None
self.projector_full = None
self.currently_selected = 0
self.resize(800, 400)
def __init__(self,
parent=None,
signalManager=None,
name="Logistic regression"):
OWWidget.__init__(self,
parent,
signalManager,
name,
wantMainArea=0,
resizingEnabled=0)
self.inputs = [("Data", ExampleTable, self.setData),
("Preprocess", PreprocessedLearner,
self.setPreprocessor)]
self.outputs = [("Learner", orange.Learner),
("Classifier", orange.Classifier), ("Features", list)]
self.regularizations = [
Orange.classification.logreg.LibLinearLogRegLearner.L2R_LR,
Orange.classification.logreg.LibLinearLogRegLearner.L1R_LR
]
self.regularizationsStr = [
"L2 (squared weights)", "L1 (absolute weights)"
]
self.name = "Logistic regression"
self.normalization = True
self.C = 1.
self.regularization = 0
self.data = None
self.preprocessor = None
self.loadSettings()
OWGUI.lineEdit(
self.controlArea,
self,
'name',
box='Learner/Classifier Name',
tooltip=
'Name to be used by other widgets to identify your learner/classifier.'
)
OWGUI.separator(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Regularization")
self.regularizationCombo = OWGUI.comboBox(
box, self, "regularization", items=self.regularizationsStr)
cset = OWGUI.doubleSpin(
box,
self,
"C",
0.01,
512.0,
0.1,
decimals=2,
addToLayout=True,
label="Training error cost (C)",
alignment=Qt.AlignRight,
tooltip=
"Weight of log-loss term (higher C means better fit on the training data)."
),
OWGUI.separator(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Preprocessing")
OWGUI.checkBox(box,
self,
"normalization",
label="Normalize data",
tooltip="Normalize data before learning.")
OWGUI.separator(self.controlArea)
applyButton = OWGUI.button(self.controlArea,
self,
"&Apply",
callback=self.applyLearner,
default=True)
OWGUI.rubber(self.controlArea)
#self.adjustSize()
self.applyLearner()
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self,
parent,
signalManager,
"CN2",
wantMainArea=0,
resizingEnabled=0)
self.inputs = [("Data", ExampleTable, self.dataset),
("Preprocess", PreprocessedLearner,
self.setPreprocessor)]
self.outputs = [("Learner", orange.Learner),
("Classifier", orange.Classifier),
("Unordered CN2 Classifier",
orngCN2.CN2UnorderedClassifier)]
self.QualityButton = 0
self.CoveringButton = 0
self.Alpha = 0.05
self.stepAlpha = 0.2
self.BeamWidth = 5
self.MinCoverage = 0
self.MaxRuleLength = 0
self.useMaxRuleLength = False
self.Weight = 0.9
self.m = 2
self.name = "CN2 rules"
self.loadSettings()
self.data = None
self.preprocessor = None
##GUI
labelWidth = 150
self.learnerName = OWGUI.lineEdit(
self.controlArea,
self,
"name",
box="Learner/classifier name",
tooltip=
"Name to be used by other widgets to identify the learner/classifier"
)
#self.learnerName.setText(self.name)
OWGUI.separator(self.controlArea)
self.ruleQualityBG = OWGUI.widgetBox(self.controlArea,
"Rule quality estimation")
self.ruleQualityBG.buttons = []
OWGUI.separator(self.controlArea)
self.ruleValidationGroup = OWGUI.widgetBox(self.controlArea,
"Pre-prunning (LRS)")
OWGUI.separator(self.controlArea)
OWGUI.spin(
self.controlArea,
self,
"BeamWidth",
1,
100,
box="Beam width",
tooltip=
"The width of the search beam\n(number of rules to be specialized)"
)
OWGUI.separator(self.controlArea)
self.coveringAlgBG = OWGUI.widgetBox(self.controlArea,
"Covering algorithm")
self.coveringAlgBG.buttons = []
"""
self.ruleQualityBG=OWGUI.radioButtonsInBox(self.ruleQualityGroup, self, "QualityButton",
btnLabels=["Laplace","m-estimate","WRACC"],
box="Rule quality", callback=self.qualityButtonPressed,
tooltips=["Laplace rule evaluator", "m-estimate rule evaluator",
"WRACC rule evaluator"])
self.mSpin=Spin=OWGUI.spin(self.ruleQualityGroup, self, "m", 0, 100, label="m",
orientation="horizontal", labelWidth=labelWidth-100, tooltip="m value for m estimate rule evaluator")
"""
b1 = QRadioButton("Laplace", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b1)
g = OWGUI.widgetBox(self.ruleQualityBG, orientation="horizontal")
b2 = QRadioButton("m-estimate", g)
g.layout().addWidget(b2)
self.mSpin = OWGUI.doubleSpin(g, self, "m", 0, 100)
b3 = QRadioButton("EVC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b3)
b4 = QRadioButton("WRACC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b4)
self.ruleQualityBG.buttons = [b1, b2, b3, b4]
for i, button in enumerate([b1, b2, b3, b4]):
self.connect(button,
SIGNAL("clicked()"),
lambda v=i: self.qualityButtonPressed(v))
OWGUI.doubleSpin(
self.ruleValidationGroup,
self,
"Alpha",
0,
1,
0.001,
label="Alpha (vs. default rule)",
orientation="horizontal",
labelWidth=labelWidth,
tooltip=
"Required significance of the difference between the class distribution on all example and covered examples"
)
OWGUI.doubleSpin(
self.ruleValidationGroup,
self,
"stepAlpha",
0,
1,
0.001,
label="Stopping Alpha (vs. parent rule)",
orientation="horizontal",
labelWidth=labelWidth,
tooltip="Required significance of each specialization of a rule.")
OWGUI.spin(
self.ruleValidationGroup,
self,
"MinCoverage",
0,
100,
label="Minimum coverage",
orientation="horizontal",
labelWidth=labelWidth,
tooltip=
"Minimum number of examples a rule must\ncover (use 0 for not setting the limit)"
)
OWGUI.checkWithSpin(
self.ruleValidationGroup,
self,
"Maximal rule length",
0,
100,
"useMaxRuleLength",
"MaxRuleLength",
labelWidth=labelWidth,
tooltip=
"Maximal number of conditions in the left\npart of the rule (use 0 for don't care)"
)
"""
self.coveringAlgBG=OWGUI.radioButtonsInBox(self.coveringAlgGroup, self, "CoveringButton",
btnLabels=["Exclusive covering ","Weighted Covering"],
tooltips=["Each example will only be used once\n for the construction of a rule",
"Examples can take part in the construction\n of many rules(CN2-SD Algorithm)"],
box="Covering algorithm", callback=self.coveringAlgButtonPressed)
self.weightSpin=OWGUI.doubleSpin(self.coveringAlgGroup, self, "Weight",0, 0.95,0.05,label= "Weight",
orientation="horizontal", labelWidth=labelWidth, tooltip=
"Multiplication constant by which the weight of\nthe example will be reduced")
"""
B1 = QRadioButton("Exclusive covering", self.coveringAlgBG)
self.coveringAlgBG.layout().addWidget(B1)
g = OWGUI.widgetBox(self.coveringAlgBG, orientation="horizontal")
B2 = QRadioButton("Weighted covering", g)
g.layout().addWidget(B2)
self.coveringAlgBG.buttons = [B1, B2]
self.weightSpin = OWGUI.doubleSpin(g, self, "Weight", 0, 0.95, 0.05)
for i, button in enumerate([B1, B2]):
self.connect(button,
SIGNAL("clicked()"),
lambda v=i: self.coveringAlgButtonPressed(v))
OWGUI.separator(self.controlArea)
self.btnApply = OWGUI.button(self.controlArea,
self,
"&Apply",
callback=self.applySettings,
default=True)
self.Alpha = float(self.Alpha)
self.stepAlpha = float(self.stepAlpha)
self.Weight = float(self.Weight)
#self.ruleQualityBG.buttons[self.QualityButton].setChecked(1)
self.qualityButtonPressed(self.QualityButton)
self.coveringAlgButtonPressed(self.CoveringButton)
self.resize(100, 100)
self.setLearner()
def __init__(self, parent=None, signalManager=None, name="Sens profile"):
OWWidget.__init__(self, parent, signalManager, name, wantMainArea=False)
self.inputs = [("Sensitivity", ExampleTable, self.sensData)]
self.outputs = [("Molecule fragmetns", ExampleTable), ("Slim-based molecular profiles", ExampleTable), ("Molecule-based slim profiles", ExampleTable), ("Slim-based fragment profiles", ExampleTable), ("Fragment-based slim profiles", ExampleTable)]
self.aspectNum = 0
self.fragFreq = 0.1
self.smilesFileName = ""
self.slimsOnly = 1
self.maxTerms = 100
self.loadSettings()
OWGUI.radioButtonsInBox(self.controlArea, self, "aspectNum",
["Process", "Function", "Component"], "Aspect")
OWGUI.checkBox(self.controlArea, self, "slimsOnly", "Slims only",
tooltip="Search for slim GO terms only")
OWGUI.spin(self.controlArea, self, "maxTerms", 10, 1000, box="Number of terms",
tooltip="Maximum number of GO terms")
self.findTermsButton = OWGUI.button(self.controlArea, self, "Find terms",
tooltip="Run term search",
callback=self.findTerms)
OWGUI.button(self.controlArea, self, "Load selected terms",
tooltip="Load selected terms from a file",
callback=self.loadSelectedTerms)
OWGUI.doubleSpin(self.controlArea, self, "fragFreq", 0.05, 0.95, 0.05,
box="Fragment frequency",
tooltip="Minimum fragment frequency")
self.findFragmentsButton = OWGUI.button(self.controlArea, self, "Find fragments",
tooltip="Run fragment search",
callback=self.findFragments)
OWGUI.button(self.controlArea, self, "Load smiles",
tooltip="Load fragment SMILES descriptors from a file",
callback=self.loadSmiles)
self.commitButton = OWGUI.button(self.controlArea, self, "Commit",
callback=self.commitData)
self.findTermsButton.setEnabled(0)
self.findFragmentsButton.setEnabled(0)
self.commitButton.setEnabled(0)
self.resize(100,100)
self.fragments = []
self.sensDict = {}
self.smilesDict = {}
self.reverseSmilesDict = {}
self.data = None
self.fragmentMap = None
self.terms = None
self.ontology = None
self.annotations = None
self.selectedTerms = []
self.lastDir = os.path.expanduser("~/")
self.loadSmiles(self.smilesFileName)
self.loadSettings()
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self, parent, signalManager, "Predictions")
self.callbackDeposit = []
self.inputs = [("Examples", ExampleTable, self.setData),
("Predictors", orange.Classifier, self.setPredictor,
Multiple)]
self.outputs = [("Predictions", ExampleTable)]
self.predictors = {}
# saveble settings
self.ShowAttributeMethod = 0
self.classes = []
self.selectedClasses = []
self.loadSettings()
self.datalabel = "N/A"
self.predictorlabel = "N/A"
self.tasklabel = "N/A"
self.outvar = None # current output variable (set by the first predictor/data set send in)
self.classifications = []
self.rindx = None
self.data = None
self.verbose = 0
self.nVarImportance = 0
# GUI - Options
# Options - classification
ibox = OWGUI.widgetBox(self.controlArea, "Info")
OWGUI.label(ibox, self, "Data: %(datalabel)s")
OWGUI.label(ibox, self, "Predictors: %(predictorlabel)s")
OWGUI.label(ibox, self, "Task: %(tasklabel)s")
OWGUI.separator(ibox)
OWGUI.label(
ibox, self,
"Predictions can be viewed with the 'Data Table'\nand saved with the 'Save' widget!"
)
OWGUI.separator(self.controlArea)
self.copt = OWGUI.widgetBox(self.controlArea,
"Probabilities (classification)")
self.copt.setDisabled(1)
self.lbcls = OWGUI.listBox(self.copt,
self,
"selectedClasses",
"classes",
selectionMode=QListWidget.MultiSelection)
self.lbcls.setFixedHeight(50)
OWGUI.separator(self.controlArea)
self.VarImportanceBox = OWGUI.doubleSpin(
self.controlArea,
self,
"nVarImportance",
0,
9999999,
1,
label="Variable importance",
orientation="horizontal",
tooltip="The number of variables to report for each prediction.")
OWGUI.checkBox(
self.controlArea,
self,
'verbose',
'Verbose',
tooltip=
'Show detailed info while predicting. This will slow down the predictions process!'
)
OWGUI.separator(self.controlArea)
self.outbox = OWGUI.widgetBox(self.controlArea, "Output")
self.apply = OWGUI.button(self.outbox,
self,
"Apply",
callback=self.sendpredictions)
OWGUI.rubber(self.controlArea)
self.adjustSize()
def __init__(self, parent=None, signalManager=None, title="PCA"):
OWWidget.__init__(self, parent, signalManager, title, wantGraph=True)
self.inputs = [("Input Data", Orange.data.Table, self.set_data)]
self.outputs = [("Transformed Data", Orange.data.Table, Default),
("Eigen Vectors", Orange.data.Table)]
self.standardize = True
self.max_components = 0
self.variance_covered = 100.0
self.use_generalized_eigenvectors = False
self.auto_commit = False
self.loadSettings()
self.data = None
self.changed_flag = False
#####
# GUI
#####
grid = QGridLayout()
box = OWGUI.widgetBox(self.controlArea,
"Components Selection",
orientation=grid)
label1 = QLabel("Max components", box)
grid.addWidget(label1, 1, 0)
sb1 = OWGUI.spin(box,
self,
"max_components",
0,
1000,
tooltip="Maximum number of components",
callback=self.on_update,
addToLayout=False,
keyboardTracking=False)
self.max_components_spin = sb1.control
self.max_components_spin.setSpecialValueText("All")
grid.addWidget(sb1.control, 1, 1)
label2 = QLabel("Variance covered", box)
grid.addWidget(label2, 2, 0)
sb2 = OWGUI.doubleSpin(box,
self,
"variance_covered",
1.0,
100.0,
1.0,
tooltip="Percent of variance covered.",
callback=self.on_update,
decimals=1,
addToLayout=False,
keyboardTracking=False)
sb2.control.setSuffix("%")
grid.addWidget(sb2.control, 2, 1)
OWGUI.rubber(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Commit")
cb = OWGUI.checkBox(box, self, "auto_commit", "Commit on any change")
b = OWGUI.button(box, self, "Commit", callback=self.update_components)
OWGUI.setStopper(self, b, cb, "changed_flag", self.update_components)
self.scree_plot = ScreePlot(self)
# self.scree_plot.set_main_title("Scree Plot")
# self.scree_plot.set_show_main_title(True)
self.scree_plot.set_axis_title(owaxis.xBottom, "Principal Components")
self.scree_plot.set_show_axis_title(owaxis.xBottom, 1)
self.scree_plot.set_axis_title(owaxis.yLeft, "Proportion of Variance")
self.scree_plot.set_show_axis_title(owaxis.yLeft, 1)
self.variance_curve = self.scree_plot.add_curve(
"Variance",
Qt.red,
Qt.red,
2,
xData=[],
yData=[],
style=OWCurve.Lines,
enableLegend=True,
lineWidth=2,
autoScale=1,
x_axis_key=owaxis.xBottom,
y_axis_key=owaxis.yLeft,
)
self.cumulative_variance_curve = self.scree_plot.add_curve(
"Cumulative Variance",
Qt.darkYellow,
Qt.darkYellow,
2,
xData=[],
yData=[],
style=OWCurve.Lines,
enableLegend=True,
lineWidth=2,
autoScale=1,
x_axis_key=owaxis.xBottom,
y_axis_key=owaxis.yLeft,
)
self.mainArea.layout().addWidget(self.scree_plot)
self.connect(self.scree_plot, SIGNAL("cutoff_moved(double)"),
self.on_cutoff_moved)
self.connect(self.graphButton, SIGNAL("clicked()"),
self.scree_plot.save_to_file)
self.components = None
self.variances = None
self.variances_sum = None
self.projector_full = None
self.currently_selected = 0
self.resize(800, 400)
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self, parent, signalManager, "CN2", wantMainArea=False, resizingEnabled=False)
self.inputs = [("Data", ExampleTable, self.dataset), ("Preprocess", PreprocessedLearner, self.setPreprocessor)]
self.outputs = [
("Learner", orange.Learner),
("Classifier", orange.Classifier),
("Unordered CN2 Classifier", orngCN2.CN2UnorderedClassifier),
]
self.QualityButton = 0
self.CoveringButton = 0
self.Alpha = 0.05
self.stepAlpha = 0.2
self.BeamWidth = 5
self.MinCoverage = 0
self.MaxRuleLength = 0
self.useMaxRuleLength = False
self.Weight = 0.9
self.m = 2
self.name = "CN2 rules"
self.loadSettings()
self.data = None
self.preprocessor = None
##GUI
labelWidth = 150
self.learnerName = OWGUI.lineEdit(
self.controlArea,
self,
"name",
box="Learner/classifier name",
tooltip="Name to be used by other widgets to identify the learner/classifier",
)
# self.learnerName.setText(self.name)
OWGUI.separator(self.controlArea)
self.ruleQualityBG = OWGUI.widgetBox(self.controlArea, "Rule quality estimation")
self.ruleQualityBG.buttons = []
OWGUI.separator(self.controlArea)
self.ruleValidationGroup = OWGUI.widgetBox(self.controlArea, "Pre-prunning (LRS)")
OWGUI.separator(self.controlArea)
OWGUI.spin(
self.controlArea,
self,
"BeamWidth",
1,
100,
box="Beam width",
tooltip="The width of the search beam\n(number of rules to be specialized)",
)
OWGUI.separator(self.controlArea)
self.coveringAlgBG = OWGUI.widgetBox(self.controlArea, "Covering algorithm")
self.coveringAlgBG.buttons = []
b1 = QRadioButton("Laplace", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b1)
g = OWGUI.widgetBox(self.ruleQualityBG, orientation="horizontal")
b2 = QRadioButton("m-estimate", g)
g.layout().addWidget(b2)
self.mSpin = OWGUI.doubleSpin(g, self, "m", 0, 100)
b3 = QRadioButton("EVC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b3)
b4 = QRadioButton("WRACC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b4)
self.ruleQualityBG.buttons = [b1, b2, b3, b4]
for i, button in enumerate([b1, b2, b3, b4]):
self.connect(button, SIGNAL("clicked()"), lambda v=i: self.qualityButtonPressed(v))
form = QFormLayout(
labelAlignment=Qt.AlignLeft, formAlignment=Qt.AlignLeft, fieldGrowthPolicy=QFormLayout.AllNonFixedFieldsGrow
)
self.ruleValidationGroup.layout().addLayout(form)
alpha_spin = OWGUI.doubleSpin(
self.ruleValidationGroup,
self,
"Alpha",
0,
1,
0.001,
tooltip="Required significance of the difference between the "
+ "class distribution on all examples and covered examples",
)
step_alpha_spin = OWGUI.doubleSpin(
self.ruleValidationGroup,
self,
"stepAlpha",
0,
1,
0.001,
tooltip="Required significance of each specialization of a rule.",
)
min_coverage_spin = OWGUI.spin(
self.ruleValidationGroup,
self,
"MinCoverage",
0,
100,
tooltip="Minimum number of examples a rule must cover " + "(use 0 for not setting the limit)",
)
min_coverage_spin.setSpecialValueText("Unlimited")
# Check box needs to be in alayout for the form layout to center it
# in the vertical direction.
max_rule_box = OWGUI.widgetBox(self.ruleValidationGroup, "")
max_rule_cb = OWGUI.checkBox(max_rule_box, self, "useMaxRuleLength", "Maximal rule length")
max_rule_spin = OWGUI.spin(
self.ruleValidationGroup,
self,
"MaxRuleLength",
1,
100,
tooltip="Maximal number of conditions in the left " + "part of the rule",
)
max_rule_cb.disables += [max_rule_spin]
max_rule_cb.makeConsistent()
form.addRow("Alpha (vs. default rule)", alpha_spin)
form.addRow("Stopping Alpha (vs. parent rule)", step_alpha_spin)
form.addRow("Minimum coverage", min_coverage_spin)
form.addRow(max_rule_box, max_rule_spin)
B1 = QRadioButton("Exclusive covering", self.coveringAlgBG)
self.coveringAlgBG.layout().addWidget(B1)
g = OWGUI.widgetBox(self.coveringAlgBG, orientation="horizontal")
B2 = QRadioButton("Weighted covering", g)
g.layout().addWidget(B2)
self.coveringAlgBG.buttons = [B1, B2]
self.weightSpin = OWGUI.doubleSpin(g, self, "Weight", 0, 0.95, 0.05)
for i, button in enumerate([B1, B2]):
self.connect(button, SIGNAL("clicked()"), lambda v=i: self.coveringAlgButtonPressed(v))
OWGUI.separator(self.controlArea)
self.btnApply = OWGUI.button(self.controlArea, self, "&Apply", callback=self.applySettings, default=True)
self.Alpha = float(self.Alpha)
self.stepAlpha = float(self.stepAlpha)
self.Weight = float(self.Weight)
# self.ruleQualityBG.buttons[self.QualityButton].setChecked(1)
self.qualityButtonPressed(self.QualityButton)
self.coveringAlgButtonPressed(self.CoveringButton)
self.resize(100, 100)
self.setLearner()
def __init__(self, parent=None, signalManager = None, name='kNN'):
OWWidget.__init__(self, parent, signalManager, name, wantMainArea = 0, resizingEnabled = 0)
self.callbackDeposit = []
self.inputs = [("Examples", ExampleTable, self.setData)]
self.outputs = [("Learner", orange.Learner),("Classifier", orange.Classifier)]
self.boostType = [("Discrete Boosting","DISCRETE"),
("Real Boosting","REAL"),
("Logit Boosting","LOGIT"),
("Gentle Boosting","GENTLE")
]
self.splitCrit = [("Default","DEFAULT"),
("Gini","GINI"),
("MissClass","MISCLASS"),
("Squared Error","SQERR")
]
self.currentSplitCrits = []
# Settings
self.name = 'CvBoost'
self.classifier = None
self.learner = None
self.modelFile = os.path.join(os.getcwd(),"Boost.model")
#CVBOOSTTYPE = { "DISCRETE":0, "REAL":1, "LOGIT":2, "GENTLE":3 }
#CVBOOSTSPLITCRIT{ "DEFAULT":0, "GINI":1, "MISCLASS":3, "SQERR":4 }
#CVBOOSTDEFAULTDICT = {"boost_type":"DISCRETE","weak_count":100,"split_criteria":"DEFAULT","weight_trim_rate":0.95, "max_depth":1, "use_surrogates":True, "priors":None}
for par in ("boost_type","weak_count","split_criteria","weight_trim_rate", "max_depth", "use_surrogates","priors"):
setattr(self, par, AZOC.CVBOOSTDEFAULTDICT[par])
self.data = None # input data set
OWGUI.lineEdit(self.controlArea, self, 'name', box='Learner/Classifier Name', \
tooltip='Name to be used by other widgets to identify the learner/classifier.')
OWGUI.separator(self.controlArea)
pars = OWGUI.widgetBox(self.controlArea, "Parameters")
self.bType = [x[1] for x in self.boostType].index(self.boost_type)
self.splitC = 0
OWGUI.comboBox(pars, self, "bType", items = [x[0] for x in self.boostType], label = "Boost Type", orientation="horizontal", valueType = str,callback= self.updateSplitCritCombo)
self.splitCombo = OWGUI.comboBox(pars, self, "splitC", items = [x[0] for x in self.splitCrit], label = "Split Criteria", orientation="horizontal", valueType = str, tooltip = "Node splitting criteria.")
self.updateSplitCritCombo()
if self.split_criteria not in [x[1] for x in self.currentSplitCrits]:
print "ERROR: Invalid CvBoost default parameters in the AZOrangeConfig.py file"
self.splitC = [x[1] for x in self.currentSplitCrits].index(self.split_criteria)
OWGUI.spin(pars, self, "weak_count", 0, 1000, 1, None, "Weak Count ", orientation="horizontal", tooltip = "The number of trees used.")
wtrBox = OWGUI.doubleSpin(pars, self, "weight_trim_rate", 0.0, 1.0, 0.01, label="Weight Trim Rate", orientation="horizontal", tooltip = "Removal of well classified examples. Used only for computational efficiency. The greater the number, the fewer the number of examples removed.")
wtrBox.control.setDecimals(2)
OWGUI.spin(pars, self, "max_depth", 1, 1000, 1, None, "Max Depth ", orientation="horizontal", tooltip = "Maximal branching depth")
OWGUI.checkBox(pars, self, "use_surrogates", "Use Surrogates", tooltip = "Use surrogate nodes for missing values")
OWGUI.separator(self.controlArea)
self.setLearner()
OWGUI.button(self.controlArea, self, "&Apply", callback = self.setLearner, disabled=0)
# Get desired location of model file
boxFile = OWGUI.widgetBox(self.controlArea, "Path for saving Model", addSpace = True, orientation=0)
L1 = OWGUI.lineEdit(boxFile, self, "modelFile", labelWidth=80, orientation = "horizontal", \
tooltip = "Once a model is created (connect this widget with a data widget), \nit can be saved by giving a \
file name here and clicking the save button.\nPlease observe that model names should not contain an extention.")
L1.setMinimumWidth(200)
button = OWGUI.button(boxFile, self, '...', callback = self.browseFile, disabled=0,tooltip = "Choose the dir where to save. After chosen, add a name for the model file!")
button.setMaximumWidth(25)
# Save the model
OWGUI.button(self.controlArea, self,"&Save model to file", callback=self.saveModel)
self.adjustSize()
def __init__(self, parent=None, signalManager=None, name='kNN'):
OWWidget.__init__(self,
parent,
signalManager,
name,
wantMainArea=0,
resizingEnabled=0)
self.callbackDeposit = []
self.inputs = [("Examples", ExampleTable, self.setData)]
self.outputs = [("Learner", orange.Learner),
("Classifier", orange.Classifier)]
self.boostType = [("Discrete Boosting", "DISCRETE"),
("Real Boosting", "REAL"),
("Logit Boosting", "LOGIT"),
("Gentle Boosting", "GENTLE")]
self.splitCrit = [("Default", "DEFAULT"), ("Gini", "GINI"),
("MissClass", "MISCLASS"),
("Squared Error", "SQERR")]
self.currentSplitCrits = []
# Settings
self.name = 'CvBoost'
self.classifier = None
self.learner = None
self.modelFile = os.path.join(os.getcwd(), "Boost.model")
#CVBOOSTTYPE = { "DISCRETE":0, "REAL":1, "LOGIT":2, "GENTLE":3 }
#CVBOOSTSPLITCRIT{ "DEFAULT":0, "GINI":1, "MISCLASS":3, "SQERR":4 }
#CVBOOSTDEFAULTDICT = {"boost_type":"DISCRETE","weak_count":100,"split_criteria":"DEFAULT","weight_trim_rate":0.95, "max_depth":1, "use_surrogates":True, "priors":None}
for par in ("boost_type", "weak_count", "split_criteria",
"weight_trim_rate", "max_depth", "use_surrogates",
"priors"):
setattr(self, par, AZOC.CVBOOSTDEFAULTDICT[par])
self.data = None # input data set
OWGUI.lineEdit(self.controlArea, self, 'name', box='Learner/Classifier Name', \
tooltip='Name to be used by other widgets to identify the learner/classifier.')
OWGUI.separator(self.controlArea)
pars = OWGUI.widgetBox(self.controlArea, "Parameters")
self.bType = [x[1] for x in self.boostType].index(self.boost_type)
self.splitC = 0
OWGUI.comboBox(pars,
self,
"bType",
items=[x[0] for x in self.boostType],
label="Boost Type",
orientation="horizontal",
valueType=str,
callback=self.updateSplitCritCombo)
self.splitCombo = OWGUI.comboBox(pars,
self,
"splitC",
items=[x[0] for x in self.splitCrit],
label="Split Criteria",
orientation="horizontal",
valueType=str,
tooltip="Node splitting criteria.")
self.updateSplitCritCombo()
if self.split_criteria not in [x[1] for x in self.currentSplitCrits]:
print "ERROR: Invalid CvBoost default parameters in the AZOrangeConfig.py file"
self.splitC = [x[1] for x in self.currentSplitCrits
].index(self.split_criteria)
OWGUI.spin(pars,
self,
"weak_count",
0,
1000,
1,
None,
"Weak Count ",
orientation="horizontal",
tooltip="The number of trees used.")
wtrBox = OWGUI.doubleSpin(
pars,
self,
"weight_trim_rate",
0.0,
1.0,
0.01,
label="Weight Trim Rate",
orientation="horizontal",
tooltip=
"Removal of well classified examples. Used only for computational efficiency. The greater the number, the fewer the number of examples removed."
)
wtrBox.control.setDecimals(2)
OWGUI.spin(pars,
self,
"max_depth",
1,
1000,
1,
None,
"Max Depth ",
orientation="horizontal",
tooltip="Maximal branching depth")
OWGUI.checkBox(pars,
self,
"use_surrogates",
"Use Surrogates",
tooltip="Use surrogate nodes for missing values")
OWGUI.separator(self.controlArea)
self.setLearner()
OWGUI.button(self.controlArea,
self,
"&Apply",
callback=self.setLearner,
disabled=0)
# Get desired location of model file
boxFile = OWGUI.widgetBox(self.controlArea,
"Path for saving Model",
addSpace=True,
orientation=0)
L1 = OWGUI.lineEdit(boxFile, self, "modelFile", labelWidth=80, orientation = "horizontal", \
tooltip = "Once a model is created (connect this widget with a data widget), \nit can be saved by giving a \
file name here and clicking the save button.\nPlease observe that model names should not contain an extention." )
L1.setMinimumWidth(200)
button = OWGUI.button(
boxFile,
self,
'...',
callback=self.browseFile,
disabled=0,
tooltip=
"Choose the dir where to save. After chosen, add a name for the model file!"
)
button.setMaximumWidth(25)
# Save the model
OWGUI.button(self.controlArea,
self,
"&Save model to file",
callback=self.saveModel)
self.adjustSize()
def __init__(self, parent=None, signalManager=None, title="Earth"):
OWWidget.__init__(self,
parent,
signalManager,
title,
wantMainArea=False)
self.inputs = [("Data", Orange.data.Table, self.set_data),
("Preprocessor", PreprocessedLearner,
self.set_preprocessor)]
self.outputs = [("Learner", earth.EarthLearner),
("Predictor", earth.EarthClassifier)]
self.name = "Earth Learner"
self.degree = 1
self.terms = 21
self.penalty = 2
self.loadSettings()
#####
# GUI
#####
OWGUI.lineEdit(self.controlArea,
self,
"name",
box="Learner/Classifier Name",
tooltip="Name for the learner/predictor")
box = OWGUI.widgetBox(self.controlArea, "Forward Pass", addSpace=True)
OWGUI.spin(
box,
self,
"degree",
1,
3,
step=1,
label="Max. term degree",
tooltip=
"Maximum degree of the terms derived (number of hinge functions).")
s = OWGUI.spin(
box,
self,
"terms",
1,
200,
step=1,
label="Max. terms",
tooltip="Maximum number of terms derived in the forward pass.")
s.control.setSpecialValueText("Automatic")
box = OWGUI.widgetBox(self.controlArea, "Pruning Pass", addSpace=True)
OWGUI.doubleSpin(box,
self,
"penalty",
min=0.0,
max=10.0,
step=0.25,
label="Knot penalty")
OWGUI.button(self.controlArea, self, "&Apply", callback=self.apply)
self.data = None
self.preprocessor = None
self.resize(300, 200)
self.apply()
def __init__(self, parent=None, signalManager=None, title="PCA"):
OWWidget.__init__(self, parent, signalManager, title, wantGraph=True)
self.inputs = [("Input Data", Orange.data.Table, self.set_data)]
self.outputs = [("Transformed Data", Orange.data.Table, Default),
("Eigen Vectors", Orange.data.Table)]
self.standardize = True
self.max_components = 0
self.variance_covered = 100.0
self.use_generalized_eigenvectors = False
self.auto_commit = False
self.loadSettings()
self.data = None
self.changed_flag = False
#####
# GUI
#####
grid = QGridLayout()
box = OWGUI.widgetBox(self.controlArea, "Components Selection",
orientation=grid)
label1 = QLabel("Max components", box)
grid.addWidget(label1, 1, 0)
sb1 = OWGUI.spin(box, self, "max_components", 0, 1000,
tooltip="Maximum number of components",
callback=self.on_update,
addToLayout=False,
keyboardTracking=False
)
self.max_components_spin = sb1.control
self.max_components_spin.setSpecialValueText("All")
grid.addWidget(sb1.control, 1, 1)
label2 = QLabel("Variance covered", box)
grid.addWidget(label2, 2, 0)
sb2 = OWGUI.doubleSpin(box, self, "variance_covered", 1.0, 100.0, 1.0,
tooltip="Percent of variance covered.",
callback=self.on_update,
decimals=1,
addToLayout=False,
keyboardTracking=False
)
sb2.control.setSuffix("%")
grid.addWidget(sb2.control, 2, 1)
OWGUI.rubber(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Commit")
cb = OWGUI.checkBox(box, self, "auto_commit", "Commit on any change")
b = OWGUI.button(box, self, "Commit",
callback=self.update_components)
OWGUI.setStopper(self, b, cb, "changed_flag", self.update_components)
self.plot = Graph()
canvas = self.plot.canvas()
canvas.setFrameStyle(QFrame.StyledPanel)
self.mainArea.layout().addWidget(self.plot)
self.plot.setAxisTitle(QwtPlot.yLeft, "Proportion of Variance")
self.plot.setAxisTitle(QwtPlot.xBottom, "Principal Components")
self.plot.setAxisScale(QwtPlot.yLeft, 0.0, 1.0)
self.plot.enableGridXB(True)
self.plot.enableGridYL(True)
self.plot.setGridColor(Qt.lightGray)
self.variance_curve = plot_curve(
"Variance",
pen=QPen(Qt.red, 2),
symbol=QwtSymbol.NoSymbol,
xaxis=QwtPlot.xBottom,
yaxis=QwtPlot.yLeft
)
self.cumulative_variance_curve = plot_curve(
"Cumulative Variance",
pen=QPen(Qt.darkYellow, 2),
symbol=QwtSymbol.NoSymbol,
xaxis=QwtPlot.xBottom,
yaxis=QwtPlot.yLeft
)
self.variance_curve.attach(self.plot)
self.cumulative_variance_curve.attach(self.plot)
self.selection_tool = CutoffControler(parent=self.plot.canvas())
self.selection_tool.cutoffMoved.connect(self.on_cutoff_moved)
self.graphButton.clicked.connect(self.saveToFile)
self.components = None
self.variances = None
self.variances_sum = None
self.projector_full = None
self.currently_selected = 0
self.resize(800, 400)
def __init__(self, parent=None, signalManager=None, title="Neural Network"):
OWWidget.__init__(self, parent, signalManager, title, wantMainArea=False, resizingEnabled=False)
self.inputs = [
("Data", Orange.data.Table, self.set_data),
("Preprocess", PreprocessedLearner, self.set_preprocessor),
]
self.outputs = [("Learner", Orange.classification.Learner), ("Classifier", Orange.classification.Classifier)]
self.name = "Neural Network"
self.n_mid = 20
self.reg_fact = 1
self.max_iter = 300
self.normalize = True
self.loadSettings()
box = OWGUI.widgetBox(self.controlArea, "Name", addSpace=True)
OWGUI.lineEdit(box, self, "name")
box = OWGUI.widgetBox(self.controlArea, "Settings", addSpace=True)
form = QFormLayout(
spacing=8,
formAlignment=Qt.AlignLeft,
labelAlignment=Qt.AlignLeft,
fieldGrowthPolicy=QFormLayout.AllNonFixedFieldsGrow,
)
box.layout().addLayout(form)
form.addRow(
"Hidden layer neurons",
OWGUI.spin(
box, self, "n_mid", 2, 10000, 1, tooltip="Number of neurons in the hidden layer.", addToLayout=False
),
)
form.addRow("Regularization factor", OWGUI.doubleSpin(box, self, "reg_fact", 0.1, 10.0, 0.1, addToLayout=False))
form.addRow(
"Max iterations",
OWGUI.spin(
box,
self,
"max_iter",
100,
10000,
1,
tooltip="Maximal number of optimization iterations.",
addToLayout=False,
),
)
OWGUI.checkBox(box, self, "normalize", "Normalize the data")
OWGUI.button(
self.controlArea,
self,
"&Apply",
callback=self.apply,
tooltip="Create the learner and apply it on input data.",
autoDefault=True,
)
self.data = None
self.preprocessor = None
self.apply()
def __init__(self, parent=None, signalManager=None, title="PCA"):
OWWidget.__init__(self, parent, signalManager, title, wantGraph=True)
self.inputs = [("Input Data", Orange.data.Table, self.set_data)]
self.outputs = [("Transformed Data", Orange.data.Table, Default),
("Eigen Vectors", Orange.data.Table)]
self.standardize = True
self.max_components = 0
self.variance_covered = 100.0
self.use_generalized_eigenvectors = False
self.auto_commit = False
self.loadSettings()
self.data = None
self.changed_flag = False
#####
# GUI
#####
grid = QGridLayout()
box = OWGUI.widgetBox(self.controlArea, "Components Selection",
orientation=grid)
label1 = QLabel("Max components", box)
grid.addWidget(label1, 1, 0)
sb1 = OWGUI.spin(box, self, "max_components", 0, 1000,
tooltip="Maximum number of components",
callback=self.on_update,
addToLayout=False,
keyboardTracking=False
)
self.max_components_spin = sb1.control
self.max_components_spin.setSpecialValueText("All")
grid.addWidget(sb1.control, 1, 1)
label2 = QLabel("Variance covered", box)
grid.addWidget(label2, 2, 0)
sb2 = OWGUI.doubleSpin(box, self, "variance_covered", 1.0, 100.0, 1.0,
tooltip="Percent of variance covered.",
callback=self.on_update,
decimals=1,
addToLayout=False,
keyboardTracking=False
)
sb2.control.setSuffix("%")
grid.addWidget(sb2.control, 2, 1)
OWGUI.rubber(self.controlArea)
box = OWGUI.widgetBox(self.controlArea, "Commit")
cb = OWGUI.checkBox(box, self, "auto_commit", "Commit on any change")
b = OWGUI.button(box, self, "Commit",
callback=self.update_components)
OWGUI.setStopper(self, b, cb, "changed_flag", self.update_components)
self.scree_plot = ScreePlot(self)
# self.scree_plot.set_main_title("Scree Plot")
# self.scree_plot.set_show_main_title(True)
self.scree_plot.set_axis_title(owaxis.xBottom, "Principal Components")
self.scree_plot.set_show_axis_title(owaxis.xBottom, 1)
self.scree_plot.set_axis_title(owaxis.yLeft, "Proportion of Variance")
self.scree_plot.set_show_axis_title(owaxis.yLeft, 1)
self.variance_curve = self.scree_plot.add_curve(
"Variance",
Qt.red, Qt.red, 2,
xData=[],
yData=[],
style=OWCurve.Lines,
enableLegend=True,
lineWidth=2,
autoScale=1,
x_axis_key=owaxis.xBottom,
y_axis_key=owaxis.yLeft,
)
self.cumulative_variance_curve = self.scree_plot.add_curve(
"Cumulative Variance",
Qt.darkYellow, Qt.darkYellow, 2,
xData=[],
yData=[],
style=OWCurve.Lines,
enableLegend=True,
lineWidth=2,
autoScale=1,
x_axis_key=owaxis.xBottom,
y_axis_key=owaxis.yLeft,
)
self.mainArea.layout().addWidget(self.scree_plot)
self.connect(self.scree_plot,
SIGNAL("cutoff_moved(double)"),
self.on_cutoff_moved
)
self.connect(self.graphButton,
SIGNAL("clicked()"),
self.scree_plot.save_to_file)
self.components = None
self.variances = None
self.variances_sum = None
self.projector_full = None
self.currently_selected = 0
self.resize(800, 400)
def __init__(self, parent=None, signalManager=None, name="SVM"):
OWWidget.__init__(self, parent, signalManager, name, wantMainArea = 0, resizingEnabled = 0)
self.inputs = [("Data", ExampleTable, self.setData),
("Preprocess", PreprocessedLearner, self.setPreprocessor)]
self.outputs = [("Learner", orange.Learner, Default),
("Classifier", orange.Classifier, Default),
("Support Vectors", ExampleTable)]
self.kernel_type = 2
self.gamma = 0.0
self.coef0 = 0.0
self.degree = 3
self.C = 1.0
self.p = 0.1
self.eps = 1e-3
self.nu = 0.5
self.shrinking = 1
self.probability=1
self.useNu=0
self.nomogram=0
self.normalization=1
self.data = None
self.selFlag=False
self.preprocessor = None
self.name="SVM"
OWGUI.lineEdit(self.controlArea, self, 'name',
box='Learner/Classifier Name',
tooltip='Name to be used by other widgets to identify your learner/classifier.')
OWGUI.separator(self.controlArea)
b = OWGUI.radioButtonsInBox(self.controlArea, self, "useNu", [],
box="SVM Type",
orientation = QGridLayout(),
addSpace=True)
b.layout().addWidget(OWGUI.appendRadioButton(b, self, "useNu", "C-SVM",
addToLayout=False),
0, 0, Qt.AlignLeft)
b.layout().addWidget(QLabel("Cost (C)", b), 0, 1, Qt.AlignRight)
b.layout().addWidget(OWGUI.doubleSpin(b, self, "C", 0.1, 512.0, 0.1,
decimals=2,
addToLayout=False,
callback=lambda *x: self.setType(0),
alignment=Qt.AlignRight,
tooltip= "Cost for a mis-classified training instance."),
0, 2)
b.layout().addWidget(OWGUI.appendRadioButton(b, self, "useNu", u"ν-SVM",
addToLayout=False),
1, 0, Qt.AlignLeft)
b.layout().addWidget(QLabel(u"Complexity bound (\u03bd)", b), 1, 1, Qt.AlignRight)
b.layout().addWidget(OWGUI.doubleSpin(b, self, "nu", 0.05, 1.0, 0.05,
tooltip="Lower bound on the ratio of support vectors",
addToLayout=False,
callback=lambda *x: self.setType(1),
alignment=Qt.AlignRight),
1, 2)
self.kernelBox=b = OWGUI.widgetBox(self.controlArea, "Kernel")
self.kernelradio = OWGUI.radioButtonsInBox(b, self, "kernel_type",
btnLabels=[u"Linear, x∙y",
u"Polynomial, (g x∙y + c)^d",
u"RBF, exp(-g|x-y|²)",
u"Sigmoid, tanh(g x∙y + c)"],
callback=self.changeKernel)
OWGUI.separator(b)
self.gcd = OWGUI.widgetBox(b, orientation="horizontal")
self.leg = OWGUI.doubleSpin(self.gcd, self, "gamma",0.0,10.0,0.0001,
decimals=5,
label=" g: ",
orientation="horizontal",
callback=self.changeKernel,
alignment=Qt.AlignRight)
self.led = OWGUI.doubleSpin(self.gcd, self, "coef0", 0.0,10.0,0.0001,
label=" c: ",
orientation="horizontal",
callback=self.changeKernel,
alignment=Qt.AlignRight)
self.lec = OWGUI.doubleSpin(self.gcd, self, "degree", 0.0,10.0,0.5,
label=" d: ",
orientation="horizontal",
callback=self.changeKernel,
alignment=Qt.AlignRight)
OWGUI.separator(self.controlArea)
self.optionsBox=b=OWGUI.widgetBox(self.controlArea, "Options",
addSpace=True)
OWGUI.doubleSpin(b,self, "eps", 0.0005, 1.0, 0.0005,
label=u"Numerical tolerance",
labelWidth = 180,
orientation="horizontal",
tooltip="Numerical tolerance of termination criterion.",
alignment=Qt.AlignRight)
self.probBox = OWGUI.checkBox(b,self, "probability",
label="Estimate class probabilities",
tooltip="Create classifiers that support class probability estimation."
)
OWGUI.checkBox(b, self, "normalization",
label="Normalize data",
tooltip="Use data normalization")
self.paramButton=OWGUI.button(self.controlArea, self, "Automatic parameter search",
callback=self.parameterSearch,
tooltip="Automatically searches for parameters that optimize classifier accuracy",
debuggingEnabled=0)
self.paramButton.setDisabled(True)
OWGUI.button(self.controlArea, self,"&Apply",
callback=self.applySettings,
default=True)
OWGUI.rubber(self.controlArea)
self.loadSettings()
self.changeKernel()
self.searching=False
self.applySettings()
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self,
parent,
signalManager,
"CN2",
wantMainArea=False,
resizingEnabled=False)
self.inputs = [("Data", ExampleTable, self.dataset),
("Preprocess", PreprocessedLearner,
self.setPreprocessor)]
self.outputs = [("Learner", orange.Learner),
("Classifier", orange.Classifier),
("Unordered CN2 Classifier",
orngCN2.CN2UnorderedClassifier)]
self.QualityButton = 0
self.CoveringButton = 0
self.Alpha = 0.05
self.stepAlpha = 0.2
self.BeamWidth = 5
self.MinCoverage = 0
self.MaxRuleLength = 0
self.useMaxRuleLength = False
self.Weight = 0.9
self.m = 2
self.name = "CN2 rules"
self.loadSettings()
self.data = None
self.preprocessor = None
##GUI
labelWidth = 150
self.learnerName = OWGUI.lineEdit(
self.controlArea,
self,
"name",
box="Learner/classifier name",
tooltip=
"Name to be used by other widgets to identify the learner/classifier"
)
#self.learnerName.setText(self.name)
OWGUI.separator(self.controlArea)
self.ruleQualityBG = OWGUI.widgetBox(self.controlArea,
"Rule quality estimation")
self.ruleQualityBG.buttons = []
OWGUI.separator(self.controlArea)
self.ruleValidationGroup = OWGUI.widgetBox(self.controlArea,
"Pre-prunning (LRS)")
OWGUI.separator(self.controlArea)
OWGUI.spin(
self.controlArea,
self,
"BeamWidth",
1,
100,
box="Beam width",
tooltip=
"The width of the search beam\n(number of rules to be specialized)"
)
OWGUI.separator(self.controlArea)
self.coveringAlgBG = OWGUI.widgetBox(self.controlArea,
"Covering algorithm")
self.coveringAlgBG.buttons = []
b1 = QRadioButton("Laplace", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b1)
g = OWGUI.widgetBox(self.ruleQualityBG, orientation="horizontal")
b2 = QRadioButton("m-estimate", g)
g.layout().addWidget(b2)
self.mSpin = OWGUI.doubleSpin(g, self, "m", 0, 100)
b3 = QRadioButton("EVC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b3)
b4 = QRadioButton("WRACC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b4)
self.ruleQualityBG.buttons = [b1, b2, b3, b4]
for i, button in enumerate([b1, b2, b3, b4]):
self.connect(button,
SIGNAL("clicked()"),
lambda v=i: self.qualityButtonPressed(v))
form = QFormLayout(labelAlignment=Qt.AlignLeft,
formAlignment=Qt.AlignLeft,
fieldGrowthPolicy=QFormLayout.AllNonFixedFieldsGrow)
self.ruleValidationGroup.layout().addLayout(form)
alpha_spin = OWGUI.doubleSpin(
self.ruleValidationGroup,
self,
"Alpha",
0,
1,
0.001,
tooltip="Required significance of the difference between the " +
"class distribution on all examples and covered examples")
step_alpha_spin = OWGUI.doubleSpin(
self.ruleValidationGroup,
self,
"stepAlpha",
0,
1,
0.001,
tooltip="Required significance of each specialization of a rule.")
min_coverage_spin = OWGUI.spin(
self.ruleValidationGroup,
self,
"MinCoverage",
0,
100,
tooltip="Minimum number of examples a rule must cover " +
"(use 0 for not setting the limit)")
min_coverage_spin.setSpecialValueText("Unlimited")
# Check box needs to be in alayout for the form layout to center it
# in the vertical direction.
max_rule_box = OWGUI.widgetBox(self.ruleValidationGroup, "")
max_rule_cb = OWGUI.checkBox(max_rule_box, self, "useMaxRuleLength",
"Maximal rule length")
max_rule_spin = OWGUI.spin(
self.ruleValidationGroup,
self,
"MaxRuleLength",
1,
100,
tooltip="Maximal number of conditions in the left " +
"part of the rule")
max_rule_cb.disables += [max_rule_spin]
max_rule_cb.makeConsistent()
form.addRow("Alpha (vs. default rule)", alpha_spin)
form.addRow("Stopping Alpha (vs. parent rule)", step_alpha_spin)
form.addRow("Minimum coverage", min_coverage_spin)
form.addRow(max_rule_box, max_rule_spin)
B1 = QRadioButton("Exclusive covering", self.coveringAlgBG)
self.coveringAlgBG.layout().addWidget(B1)
g = OWGUI.widgetBox(self.coveringAlgBG, orientation="horizontal")
B2 = QRadioButton("Weighted covering", g)
g.layout().addWidget(B2)
self.coveringAlgBG.buttons = [B1, B2]
self.weightSpin = OWGUI.doubleSpin(g, self, "Weight", 0, 0.95, 0.05)
for i, button in enumerate([B1, B2]):
self.connect(button,
SIGNAL("clicked()"),
lambda v=i: self.coveringAlgButtonPressed(v))
OWGUI.separator(self.controlArea)
self.btnApply = OWGUI.button(self.controlArea,
self,
"&Apply",
callback=self.applySettings,
default=True)
self.Alpha = float(self.Alpha)
self.stepAlpha = float(self.stepAlpha)
self.Weight = float(self.Weight)
#self.ruleQualityBG.buttons[self.QualityButton].setChecked(1)
self.qualityButtonPressed(self.QualityButton)
self.coveringAlgButtonPressed(self.CoveringButton)
self.resize(100, 100)
self.setLearner()
def __init__(self, parent=None, signalManager=None):
OWWidget.__init__(self,parent,signalManager,"CN2", wantMainArea = 0, resizingEnabled = 0)
self.inputs = [("Data", ExampleTable, self.dataset), ("Preprocess", PreprocessedLearner, self.setPreprocessor)]
self.outputs = [("Learner", orange.Learner),("Classifier",orange.Classifier),("Unordered CN2 Classifier", orngCN2.CN2UnorderedClassifier)]
self.QualityButton = 0
self.CoveringButton = 0
self.Alpha = 0.05
self.stepAlpha = 0.2
self.BeamWidth = 5
self.MinCoverage = 0
self.MaxRuleLength = 0
self.useMaxRuleLength = False
self.Weight = 0.9
self.m = 2
self.name = "CN2 rules"
self.loadSettings()
self.data=None
self.preprocessor = None
##GUI
labelWidth = 150
self.learnerName = OWGUI.lineEdit(self.controlArea, self, "name", box="Learner/classifier name", tooltip="Name to be used by other widgets to identify the learner/classifier")
#self.learnerName.setText(self.name)
OWGUI.separator(self.controlArea)
self.ruleQualityBG = OWGUI.widgetBox(self.controlArea, "Rule quality estimation")
self.ruleQualityBG.buttons = []
OWGUI.separator(self.controlArea)
self.ruleValidationGroup = OWGUI.widgetBox(self.controlArea, "Pre-prunning (LRS)")
OWGUI.separator(self.controlArea)
OWGUI.spin(self.controlArea, self, "BeamWidth", 1, 100, box="Beam width", tooltip="The width of the search beam\n(number of rules to be specialized)")
OWGUI.separator(self.controlArea)
self.coveringAlgBG = OWGUI.widgetBox(self.controlArea, "Covering algorithm")
self.coveringAlgBG.buttons = []
"""
self.ruleQualityBG=OWGUI.radioButtonsInBox(self.ruleQualityGroup, self, "QualityButton",
btnLabels=["Laplace","m-estimate","WRACC"],
box="Rule quality", callback=self.qualityButtonPressed,
tooltips=["Laplace rule evaluator", "m-estimate rule evaluator",
"WRACC rule evaluator"])
self.mSpin=Spin=OWGUI.spin(self.ruleQualityGroup, self, "m", 0, 100, label="m",
orientation="horizontal", labelWidth=labelWidth-100, tooltip="m value for m estimate rule evaluator")
"""
b1 = QRadioButton("Laplace", self.ruleQualityBG); self.ruleQualityBG.layout().addWidget(b1)
g = OWGUI.widgetBox(self.ruleQualityBG, orientation = "horizontal");
b2 = QRadioButton("m-estimate", g)
g.layout().addWidget(b2)
self.mSpin = OWGUI.doubleSpin(g,self,"m",0,100)
b3 = QRadioButton("EVC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b3)
b4 = QRadioButton("WRACC", self.ruleQualityBG)
self.ruleQualityBG.layout().addWidget(b4)
self.ruleQualityBG.buttons = [b1, b2, b3, b4]
for i, button in enumerate([b1, b2, b3, b4]):
self.connect(button, SIGNAL("clicked()"), lambda v=i: self.qualityButtonPressed(v))
OWGUI.doubleSpin(self.ruleValidationGroup, self, "Alpha", 0, 1,0.001, label="Alpha (vs. default rule)",
orientation="horizontal", labelWidth=labelWidth,
tooltip="Required significance of the difference between the class distribution on all example and covered examples")
OWGUI.doubleSpin(self.ruleValidationGroup, self, "stepAlpha", 0, 1,0.001, label="Stopping Alpha (vs. parent rule)",
orientation="horizontal", labelWidth=labelWidth,
tooltip="Required significance of each specialization of a rule.")
OWGUI.spin(self.ruleValidationGroup, self, "MinCoverage", 0, 100,label="Minimum coverage",
orientation="horizontal", labelWidth=labelWidth, tooltip=
"Minimum number of examples a rule must\ncover (use 0 for not setting the limit)")
OWGUI.checkWithSpin(self.ruleValidationGroup, self, "Maximal rule length", 0, 100, "useMaxRuleLength", "MaxRuleLength", labelWidth=labelWidth,
tooltip="Maximal number of conditions in the left\npart of the rule (use 0 for don't care)")
"""
self.coveringAlgBG=OWGUI.radioButtonsInBox(self.coveringAlgGroup, self, "CoveringButton",
btnLabels=["Exclusive covering ","Weighted Covering"],
tooltips=["Each example will only be used once\n for the construction of a rule",
"Examples can take part in the construction\n of many rules(CN2-SD Algorithm)"],
box="Covering algorithm", callback=self.coveringAlgButtonPressed)
self.weightSpin=OWGUI.doubleSpin(self.coveringAlgGroup, self, "Weight",0, 0.95,0.05,label= "Weight",
orientation="horizontal", labelWidth=labelWidth, tooltip=
"Multiplication constant by which the weight of\nthe example will be reduced")
"""
B1 = QRadioButton("Exclusive covering", self.coveringAlgBG); self.coveringAlgBG.layout().addWidget(B1)
g = OWGUI.widgetBox(self.coveringAlgBG, orientation = "horizontal")
B2 = QRadioButton("Weighted covering", g); g.layout().addWidget(B2)
self.coveringAlgBG.buttons = [B1, B2]
self.weightSpin=OWGUI.doubleSpin(g,self,"Weight",0,0.95,0.05)
for i, button in enumerate([B1, B2]):
self.connect(button, SIGNAL("clicked()"), lambda v=i: self.coveringAlgButtonPressed(v))
OWGUI.separator(self.controlArea)
self.btnApply = OWGUI.button(self.controlArea, self, "&Apply", callback=self.applySettings, default=True)
self.Alpha=float(self.Alpha)
self.stepAlpha=float(self.stepAlpha)
self.Weight=float(self.Weight)
#self.ruleQualityBG.buttons[self.QualityButton].setChecked(1)
self.qualityButtonPressed(self.QualityButton)
self.coveringAlgButtonPressed(self.CoveringButton)
self.resize(100,100)
self.setLearner()
def __init__(self, parent=None, signalManager=None, name="SVM"):
OWWidget.__init__(self, parent, signalManager, name)
self.inputs=[("Example Table", ExampleTable, self.setData)]
self.outputs=[("Learner", orange.Learner),("Classifier", orange.Classifier)] # ,("Support Vectors", ExampleTable)]
self.data = None
self.name="Support Vector Machine"
self.nMin = -1
self.nMax = 1
self.nClassMin = -1
self.nClassMax = 1
self.svm_typeGUI = None
self.stopCritGUI = None
self.priorsGUI = ""
#Read default parameters from AZOrangeConfig.py file
for par in ("kernel_type", "svm_type","gamma","C","p","epsC","epsR","stopCrit",\
"maxIter","nu","scaleData","scaleClass","priors","coef0","degree"):
setattr(self, par, AZOC.CVSVMDEFAULTDICT[par])
if self.svm_type in (103,104):
self.eps = self.epsR
else:
self.eps = self.epsC
self.setSVMTypeGUI(self.svm_type)
self.setStopCritGUI(self.stopCrit)
self.setPriorsGUI()#self.priors)
##scJS Added parameter needed for saving an SVM model
self.modelFile = os.path.join(os.getcwd(),"SVM.model")
##ecJS
OWGUI.lineEdit(self.controlArea, self, 'name', box='Learner/Classifier Name', \
tooltip='Name to be used by other widgets to identify your learner/classifier.<br>This should be a unique name!')
## scLC
# Do not use the one class svm (type 2).
self.svmTypeBox=b=OWGUI.widgetBox(self.controlArea,"SVM Type")
self.svmTyperadio = OWGUI.radioButtonsInBox(b, self, "svm_typeGUI", btnLabels=["C-SVC", "nu-SVC","epsilon-SVR", "nu-SVR"], callback=self.setSVMTypeLearner)
## ecLC
self.kernelBox=b=OWGUI.widgetBox(self.controlArea, "Kernel")
self.kernelradio = OWGUI.radioButtonsInBox(b, self, "kernel_type", btnLabels=["Linear, x.y", "Polynomial, (g*x.y+c)^d",
"RBF, exp(-g*(x-y).(x-y))", "Sigmoid, tanh(g*x.y+c)"], callback=self.changeKernel)
self.gcd = OWGUI.widgetBox(b, orientation="horizontal")
self.leg = OWGUI.doubleSpin(self.gcd, self, "gamma",0.0,10.0,0.0001, label="gamma: ", orientation="horizontal", callback=self.changeKernel)
self.led = OWGUI.doubleSpin(self.gcd, self, "coef0", 0.0,10.0,0.0001, label=" coef0: ", orientation="horizontal", callback=self.changeKernel)
self.lec = OWGUI.doubleSpin(self.gcd, self, "degree", 0.0,10.0,0.5, label=" degree: ", orientation="horizontal", callback=self.changeKernel)
OWGUI.separator(self.controlArea)
#self.layout=QVBoxLayout(self.mainArea)
self.optionsBox=b=OWGUI.widgetBox(self.mainArea, "Options", addSpace = True)
self.cBox = OWGUI.doubleSpin(b,self, "C", 0.0, 512.0, 0.5, label="Model complexity (C)", orientation="horizontal", tooltip="Penalty cost")
self.cBox.control.setDecimals(3)
self.pBox = OWGUI.doubleSpin(b,self, "p", 0.0, 10.0, 0.1, label="Tolerance (p)", orientation="horizontal", tooltip="zero-loss width zone")
self.pBox.control.setDecimals(3)
self.nuBox = OWGUI.doubleSpin(b, self, "nu", 0.0,1.0,0.1, label="Complexity bound (nu)", orientation="horizontal", tooltip="Upper bound on the ratio of support vectors")
self.nuBox.control.setDecimals(3)
self.priorBox = OWGUI.lineEdit(b, self, 'priorsGUI', box='Class weights', \
tooltip='Ex: POS:2 , NEG:1')
self.stopCritBox=OWGUI.widgetBox(self.optionsBox,"Stop Criteria")
self.stopCritradio = OWGUI.radioButtonsInBox(self.stopCritBox, self, "stopCritGUI", btnLabels=["Maximum number of iterations (maxIter)", "Numeric precision (eps)"], callback=self.setStopCritLearner)
OWGUI.doubleSpin(self.stopCritBox,self, "eps", 0.0, 0.5, 0.001, label="Numeric precision (eps)", orientation="horizontal")
OWGUI.doubleSpin(self.stopCritBox,self, "maxIter", 0, 100000, 1, label="Maximum number of iterations (maxIter)", orientation="horizontal")
self.scaleDataBox=OWGUI.checkBox(b, self, "scaleData", "Scale Data", tooltip="Scales the training data, and the input examples",callback = self.changeScaling)
self.scaleClassBox=OWGUI.checkBox(OWGUI.indentedBox(b), self, "scaleClass", "Scale Class variable", tooltip="Scales the class variable of training data, and the output predictions",callback = self.changeScaling)
#self.scalenMin = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nMin', 'Var min:', orientation="horizontal", tooltip='Variables scaling minimum')
#self.scalenMax = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nMax', 'Var max:', orientation="horizontal", tooltip='Variables scaling maximum')
#OWGUI.separator(OWGUI.indentedBox(b))
self.scalenClassMin = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nClassMin', 'Class min:', orientation="horizontal", tooltip='Class scaling minimum')
self.scalenClassMax = OWGUI.lineEdit(OWGUI.indentedBox(b), self, 'nClassMax', 'Class max:', orientation="horizontal", tooltip='Class scaling maximum')
OWGUI.separator(self.controlArea)
#self.paramButton=OWGUI.button(self.controlArea, self, "Automatic parameter search", callback=self.parameterSearch,
# tooltip="Automaticaly searches for parameters that optimize classifier acuracy")
self.optionsBox.adjustSize()
#self.layout.add(self.optionsBox)
#OWGUI.separator(self.controlArea)
OWGUI.button(self.controlArea, self,"&Apply", callback=self.applySettings)
##scJS Adding saving of svm models
# Get desired location of model file
boxFile = OWGUI.widgetBox(self.controlArea, "Path for saving Model", addSpace = True, orientation=0)
L1 = OWGUI.lineEdit(boxFile, self, "modelFile", labelWidth=80, orientation = "horizontal", \
tooltip = "Once a model is created (connect this widget with a data widget), \nit can be saved by giving a \
file name here and clicking the save button.")
#L1.setMinimumWidth(200)
button = OWGUI.button(boxFile, self, '...', callback = self.browseFile, disabled=0,tooltip = "Choose the dir where to save. After chosen, add a name for the model file!")
#button.setMaximumWidth(25)
# Save model
OWGUI.button(self.controlArea, self,"&Save SVM model", callback=self.saveModel)
#self.adjustSize()
#self.loadSettings()
self.setStopCritLearner()
self.setSVMTypeLearner()
self.changeKernel()
self.applySettings()
self.resize(500,400)
