def main():
iris = datasets.load_iris()
x = iris.data[50:150]
t = iris.target[50:150]
for i in range(100):
if t[i] == 2:
t[i] = -1
skf = model_selection.StratifiedKFold(n_splits=10)
avg_accuracy = 0
for train, test in skf.split(x, t):
classifier = PerceptronClassifier(4, 0.1)
classifier.learn(t[train], x[train])
accuracy = 0.0
for test_x, test_t in zip(x[test], t[test]):
test_y = classifier.classify(test_x)
if test_y == test_t:
accuracy += 1
accuracy /= len(test)
avg_accuracy += accuracy
avg_accuracy /= 10
print 'average accuracy:', avg_accuracy * 100, '%'
def basic():
print("-------------basic---------------")
x = np.array([[0,0,1],[1,1,1],[1,0,1],[0,1,1]])
y = np.array([[0],[1],[1],[0]])
# print(x)
# print(y)
pc = PerceptronClassifier(lr=1)
# print(pc)
print(pc.fit(x, y).score(np.array([[0,0,0],[0,1,0],[1,0,0],[1,1,0]]), np.array([[0],[0],[1],[1]])))
print(pc)
def __init__(self):
iris = datasets.load_iris()
x = iris.data[0:150]
t = iris.target[0:150]
for i in range(150):
if t[i] == 0:
t[i] = -1
if t[i] == 2:
t[i] = 1
self.classifier = PerceptronClassifier(4, 0.1)
self.classifier.learn(t, x)
def test_01_from_file_02(self):
"""Verify that a classifier can be constructed from a file with weights."""
classifier = PerceptronClassifier({'good': 1, 'terrible': -1})
classifier.save('data/ex05_from_file_test.model')
classifier = PerceptronClassifier.from_file(
'data/ex05_from_file_test.model')
if classifier is None:
self.fail(
msg=
"Constructing classifier from file failed: from_file returned None"
)
expected_weights = {'good': 1, 'terrible': -1}
self.assertEqual(classifier.weights, expected_weights)
def separable():
print("----------------separable-----------------------")
mat = Arff("./separableIsSquare.arff", label_count=1)
np_mat = mat.data
data = mat[:, :-1]
labels = mat[:, -1].reshape(-1, 1)
print(data[:, 1])
print(labels)
### Make the Classifier #####
P3Class = None
for lr in range(10, 0, -1):
P3Class = PerceptronClassifier(lr=0.1*lr, shuffle=False)
P3Class.fit(data, labels, standard_weight_value=None)
Accuracy = P3Class.score(data, labels)
print("Learning Rate = ", 0.1*lr)
print("Accuracy = [{:.2f}]".format(Accuracy))
print("Epochs = ", P3Class.get_epochs_trained())
# print(P3Class)
## could not get graphing to work in time...
# graph(data[:, 0], data[:, 1], labels=mat[:, -1])
w = P3Class.get_weights()
y = lambda x: (-w[0]/w[1])*x - (w[2]/w[1])
grapher = Grapher()
grapher.graph(data[:, 0], data[:, 1], labels=mat[:, -1], title="Separable")
grapher.add_function(y)
grapher.show("separable.svg")
def inseparable():
print("----------------Inseparable-----------------------")
mat = Arff("./impossible.arff", label_count=1)
np_mat = mat.data
data = mat[:, :-1]
labels = mat[:, -1].reshape(-1, 1)
### Make the Classifier #####
P4Class = None
for lr in range(10, 0, -1):
P4Class = PerceptronClassifier(lr=0.1*lr, deterministic=10, shuffle=False)
P4Class.fit(data, labels, standard_weight_value=None)
Accuracy = P4Class.score(data, labels)
print("Learning Rate = ", 0.1*lr)
print("Accuracy = [{:.2f}]".format(Accuracy))
print("Epochs = ", P4Class.get_epochs_trained())
w = P4Class.get_weights()
y = lambda x: (-w[0]/w[1])*x - (w[2]/w[1])
grapher = Grapher()
grapher.graph(data[:, 0], data[:, 1], labels=mat[:, -1], title="Inseparable")
grapher.add_function(y)
grapher.show("Inseparable.svg")
def runMahCode(arff, shuffle=True, determ=0, training=False, lr=.1, quiet=False):
mat = Arff(arff,label_count=1)
data = mat.data[:,0:-1]
labels = mat.data[:,-1:]
PClass = PerceptronClassifier(lr=lr,shuffle=shuffle,deterministic=determ)
Accuracy = 0.0
if (training):
X_train, y_train, X_test, y_test = PerceptronClassifier.split_training(data,labels)
PClass.fit(X_train,y_train)
Accuracy = PClass.score(X_test,y_test)
else:
PClass.fit(data,labels)
Accuracy = PClass.score(data,labels)
if not quiet:
print("Accuracy = [{:.5f}]".format(Accuracy))
print("Final Weights =",PClass.get_weights())
else:
return Accuracy
def testing(num):
trainData = samples.loadImagesFile("data/digitdata/trainingimages", num, 28, 28)
trainLabels = samples.loadLabelsFile("data/digitdata/traininglabels", num)
testData = samples.loadImagesFile("data/digitdata/testimages", 1000, 28, 28)
testLabels = samples.loadLabelsFile("data/digitdata/testlabels", 1000)
validData = samples.loadImagesFile("data/digitdata/validationimages", 1000, 28, 28)
validLabels = samples.loadLabelsFile("data/digitdata/validationlabels", 1000)
perceptron=PerceptronClassifier(trainData, trainLabels,0)
perceptron.train(trainData, trainLabels,10)
print "==================================="
print "Test Data"
guess=perceptron.classify(testData)
samples.verify(perceptron, guess, testLabels)
print "==================================="
print "Validation Data"
guess=perceptron.classify(validData)
samples.verify(perceptron,guess,validLabels)
def testing(num):
trainData = samples.loadImagesFile("data/facedata/facedatatrain", num, 60, 70)
trainLabels = samples.loadLabelsFile("data/facedata/facedatatrainlabels", num)
testData = samples.loadImagesFile("data/facedata/facedatatest", 150, 60, 70)
testLabels = samples.loadLabelsFile("data/facedata/facedatatestlabels", 151)
validData = samples.loadImagesFile("data/facedata/facedatavalidation", 301, 60, 70)
validLabels = samples.loadLabelsFile("data/facedata/facedatavalidationlabels", 301)
perceptron=PerceptronClassifier(trainData, trainLabels,0)
perceptron.train(trainData, trainLabels,10)
print "==================================="
print "Test Data"
guess=perceptron.classify(testData)
samples.verify(perceptron, guess, testLabels)
print "==================================="
print "Validation Data"
guess=perceptron.classify(validData)
samples.verify(perceptron,guess,validLabels)
def test_02_from_dataset_02(self):
"""Verify that a classifier can be constructed with initial weights for a fiven dataset."""
expected_weights = {'highly': 0, 'boring': 0, 'green': 0, 'eggs': 0}
classifier = PerceptronClassifier.from_dataset(
self.small_dataset_train_2)
if classifier is None:
self.fail(
msg=
'Constructing classifier for dataset failed: from_dataset returned None'
)
self.assertEqual(classifier.weights, expected_weights)
def test_04_update_02(self):
"""Verify that the perceptron update is performed correctly."""
classifier = PerceptronClassifier({'highly': 1, 'boring': -1})
classifier.update(self.small_instance_list_do_update[1])
expected_weigths = {'highly': 1, 'boring': 0}
self.assertEqual(classifier.weights, expected_weigths)
classifier = PerceptronClassifier({'highly': 1, 'boring': -1})
do_update = classifier.update(self.small_instance_list_no_update[1])
self.assertEqual(False, do_update)
class Perceptron:
def __init__(self):
iris = datasets.load_iris()
x = iris.data[0:150]
t = iris.target[0:150]
for i in range(150):
if t[i] == 0:
t[i] = -1
if t[i] == 2:
t[i] = 1
self.classifier = PerceptronClassifier(4, 0.1)
self.classifier.learn(t, x)
def predict(self, x):
y = self.classifier.classify(x)
if y == -1:
return 'setosa'
else:
return None
def our_avg_run(avg_num_of_run,filename):
dataset = load_dataset(filename)
ptraining_error = []
perceptron_error = []
for i in range(avg_num_of_run):
(training_set, testing_set) = split_dataset(dataset, PROBABILITY_TRAINING_SET)
testing_set = dataset
if IS_VERBOSE:
print("training set size: %s testing set size: %s num instances: %s" %
(len(training_set), len(testing_set), len(dataset)))
(train_x, train_y) = split_attribute_and_label(training_set)
(test_x, test_y) = split_attribute_and_label(testing_set)
p = PerceptronClassifier(ETA, THRESHOLD, UPPER_BOUND, False)
p.fit(train_x, train_y)
t_result_list = p.predict(train_x)
ptraining_error.append(calculate_error(train_y, t_result_list))
result_list = p.predict(test_x)
perceptron_error.append(calculate_error(test_y, result_list))
print(p.weights)
return sum(perceptron_error) / len(perceptron_error) , sum(ptraining_error) / len(ptraining_error)
def test_04_update_01(self):
"""Verify that the perceptron update is performed correctly."""
classifier = PerceptronClassifier({'highly': 1, 'boring': -1})
# Test document: ("highly", "doc25", -1)
classifier.update(self.small_instance_list_do_update[0])
expected_weigths = {'highly': 0, 'boring': -1}
self.assertEqual(classifier.weights, expected_weigths)
classifier = PerceptronClassifier({'highly': 1, 'boring': -1})
# Test document: ("boring", "doc26", 1),
do_update = classifier.update(self.small_instance_list_no_update[0])
self.assertEqual(False, do_update)
def nltk_movie_review_accuracy(num_iterations):
""" Try different number of features, and optimize number of training iterations."""
return 0, 0 # TODO: Exercise 4: remove line
(training_documents, dev_documents, test_documents) = load_reviews()
best_development_accuracy = 0.0
best_num_features = 0
best_classifier = None
best_feature_set = None
# Test different numbers of features.
for n in [100, 1000, 10000]:
print("Training with %d features..." % n)
# Training set
training_set = Dataset.from_document_collection(training_documents,
num_features=n)
# Development set
development_set = Dataset.from_document_collection(
dev_documents, feature_set=training_set.feature_set)
# Train classifier
classifier = PerceptronClassifier.from_dataset(training_set)
pass # TODO: Exercise 4: train the classifier
# Accuracies of classifier with n features
train_accuracy = classifier.test_accuracy(training_set)
development_accuracy = classifier.test_accuracy(development_set)
if development_accuracy > best_development_accuracy:
best_development_accuracy = development_accuracy
best_num_features = n
best_classifier = classifier.copy()
best_feature_set = training_set.feature_set
print(
"Best classifier with %d features: \t Train Accuracy: %.4f \t Dev Accuracy: %.4f"
% (n, train_accuracy, best_development_accuracy))
print("Best number of features: %d " % best_num_features)
print("Top features for positive class:")
print(best_classifier.features_for_class(True))
print("Top features for negative class:")
print(best_classifier.features_for_class(False))
# Compute test score for best setting.
testing_set = Dataset.from_document_collection(
test_documents, feature_set=best_feature_set)
testing_accuracy = best_classifier.test_accuracy(testing_set)
print("Test score for best setting: %.4f" % testing_accuracy)
return best_development_accuracy, testing_accuracy
def debug():
print("------------arff-------------------")
mat = Arff("../data/perceptron/debug/linsep2nonorigin.arff", label_count=1)
data = mat.data[:, 0:-1]
labels = mat.data[:, -1].reshape(-1, 1)
PClass = PerceptronClassifier(
lr=0.1, shuffle=False, deterministic=10, printIt=False)
PClass.fit(data, labels)
Accuracy = PClass.score(data, labels)
print("Accuray = [{:.2f}]".format(Accuracy))
print("Final Weights =", PClass.get_weights())
def voting():
print("--------------voting---------------------")
mat = Arff("../data/perceptron/vote.arff", label_count=1)
np_mat = mat.data
avg = []
for iteration in range(5):
print("xxxxxxxxxxx " + str(iteration) + " xxxxxxxx")
training, testing = _shuffle_split(mat.data, .3)
data = training[:, :-1]
labels = training[:, -1].reshape(-1, 1)
P5Class = PerceptronClassifier(lr=0.1, shuffle=True)
P5Class.fit(data, labels)
Accuracy = P5Class.score(data, labels)
print("Accuracy = [{:.2f}]".format(Accuracy))
print("Epochs = ", P5Class.get_epochs_trained())
tData = testing[:, :-1]
tLabels = testing[:, -1].reshape(-1, 1)
tAccuracy = P5Class.score(tData, tLabels)
print("Test Accuracy = [{:.2f}]".format(tAccuracy))
weights = P5Class.get_weights()
print(weights)
sort_weights = sorted(zip(weights, list(range(len(weights)))), key=lambda x: abs(x[0]), reverse=True)
print("sorted:\r\n", sort_weights)
scores = P5Class.getTrace().getColumns("epochScore")
print('scores', scores)
avg.append((float(scores[-2][0]) - float(scores[0][0])) / len(scores))
print('avg', avg)
grapher = Grapher()
grapher.graph(list(range(len(avg))), avg, labels=[1]*len(avg), points=False, title="Average Scores", xlabel="Iteration", ylabel="score")
grapher.show("AverageScores.svg")
def evaluation():
print("--------------arf2------------------------------")
mat = Arff("../data/perceptron/evaluation/data_banknote_authentication.arff", label_count=1)
np_mat = mat.data
data = mat[:, :-1]
labels = mat[:, -1].reshape(-1, 1)
#### Make Classifier ####
P2Class = PerceptronClassifier(lr=0.1, shuffle=False, deterministic=10)
P2Class.fit(data, labels)
Accuracy = P2Class.score(data, labels)
print("Accuray = [{:.2f}]".format(Accuracy))
print("Final Weights =", P2Class.get_weights())
def testing(num):
trainData = samples.loadImagesFile("data/digitdata/trainingimages", num, 28, 28)
trainLabels = samples.loadLabelsFile("data/digitdata/traininglabels", num)
testData = samples.loadImagesFile("data/digitdata/testimages", 1000, 28, 28)
testLabels = samples.loadLabelsFile("data/digitdata/testlabels", 1000)
validData = samples.loadImagesFile("data/digitdata/validationimages", 1000, 28, 28)
validLabels = samples.loadLabelsFile("data/digitdata/validationlabels", 1000)
perceptron=PerceptronClassifier(trainData, trainLabels,0)
perceptron.train(trainData, trainLabels,10)
print "***********************************"
print "*************Test Data*************"
guess=perceptron.classify(testData)
samples.verify(perceptron, guess, testLabels)
print "***********************************"
print "************Valid Data*************"
guess=perceptron.classify(validData)
samples.verify(perceptron,guess,validLabels)
def testing(num):
trainData = samples.loadImagesFile("data/facedata/facedatatrain", num, 60,
70)
trainLabels = samples.loadLabelsFile("data/facedata/facedatatrainlabels",
num)
testData = samples.loadImagesFile("data/facedata/facedatatest", 150, 60,
70)
testLabels = samples.loadLabelsFile("data/facedata/facedatatestlabels",
151)
validData = samples.loadImagesFile("data/facedata/facedatavalidation", 301,
60, 70)
validLabels = samples.loadLabelsFile(
"data/facedata/facedatavalidationlabels", 301)
perceptron = PerceptronClassifier(trainData, trainLabels, 0)
perceptron.train(trainData, trainLabels, 10)
print "==================================="
print "Test Data"
guess = perceptron.classify(testData)
samples.verify(perceptron, guess, testLabels)
print "==================================="
print "Validation Data"
guess = perceptron.classify(validData)
samples.verify(perceptron, guess, validLabels)
from perceptron import PerceptronClassifier
from arff import Arff
import numpy as np
mat = Arff("../data/perceptron/evaluation/data_banknote_authentication.arff",
label_count=1)
data = mat.data[:, 0:-1]
labels = mat.data[:, -1:]
PClass = PerceptronClassifier(lr=0.1, shuffle=False, deterministic=10)
PClass.fit(data, labels)
Accuracy = PClass.score(data, labels)
print("Accuray = [{:.5f}]".format(Accuracy))
print("Final Weights =", PClass.get_weights())
DIR: 'data/digitdata',
HEIGHT: 20,
WIDTH: 29,
LABEL: 10,
PIXELS: None
}
}
samples = Samples(map.get(inp).get(DIR))
dataClassifier = DataClassifier(
map.get(inp).get(HEIGHT),
map.get(inp).get(WIDTH),
map.get(inp).get(LABEL),
map.get(inp).get(PIXELS))
perceptronClassifier = PerceptronClassifier(dataClassifier.FEATURES,
dataClassifier.LABELS)
samples.readFiles()
dataset = 0
featureValueListForAllTrainingImages, actualLabelForTrainingList = dataClassifier.extractFeatures(
samples.train_lines_itr, samples.train_labelsLines_itr)
TOTALDATASET = len(featureValueListForAllTrainingImages)
INCREMENTS = int(TOTALDATASET * PERCENT_INCREMENT / 100)
PERCEPTRON_TIME = {}
while dataset < TOTALDATASET:
startTimer = time.time()
print("Training ON {0} to {1} data".format(dataset,
plt.xlabel("X: pH of Coffee")
plt.show();
# add the plot later
# # 3. Train on both sets
# Linearly Separable:
# In[9]:
data = linSep[:,:-1]
labels = linSep[:,-1:]
for lr in np.linspace(.01, 1, 10):
PClass = PerceptronClassifier(lr=lr,shuffle=True,deterministic=0)
PClass.fit(data,labels)
Accuracy = PClass.score(data,labels)
print(Accuracy)
# Because everything is centered very close to zero, it doesn't take very long for the learning rate to differentiate the two sets.
#
# We also see that the learning rate doesn't change much.
# In[10]:
data = nonLin[:,:-1]
labels = nonLin[:,-1:]
for lr in np.linspace(.01, 1, 10):
import matplotlib.pyplot as plt
from utils import DatasetGenerator
from perceptron import PerceptronClassifier
import numpy as np
discr_func = 'ellipse' # choose one of ['linear', 'ellipse', 'quadratic']
data_generator = DatasetGenerator()
X, y = data_generator.generate_separable_data(num_points=57,
discr_func=discr_func)
clf = PerceptronClassifier(discr_func=discr_func)
clf.fit(X, y)
plt.xlim(-14, 14)
plt.ylim(-14, 14)
# plot data points
plt.scatter(X[:, 0], X[:, 1], s=25, c=y)
# plot separating curve
xmin, xmax = plt.xlim()
ymin, ymax = plt.ylim()
xlist = np.linspace(xmin, xmax, 100)
ylist = np.linspace(ymin, ymax, 100)
XX, YY = np.meshgrid(xlist, ylist)
xy = np.vstack([XX.ravel(), YY.ravel()]).T
Z = clf.score(xy).reshape(XX.shape)
plt.contour(XX, YY, Z, levels=[0], colors=['r'])
plt.show()
import statistics
import time
import numpy as np
import p3_utils
from load_data import FaceData
from perceptron import PerceptronClassifier
if __name__ == '__main__':
iterations = 3
legalLabels = range(2)
face_data = FaceData("facedata")
classifier = PerceptronClassifier(legalLabels, iterations)
featureFunction = face_data.basic_feature_extractor
trainingDataList = list(map(featureFunction, face_data.face_train_imgs))
validationDataList = list(
map(featureFunction, face_data.face_validation_imgs))
testDataList = list(map(featureFunction, face_data.face_test_imgs))
n_train = len(trainingDataList)
classifier.set_weights(range(2), FaceData.FACE_DATUM_WIDTH,
FaceData.FACE_DATUM_HEIGHT)
# Conduct training and testing
percentages, runtimes = ([], [])
for n, n_samples in enumerate(
range(n_train // 10, n_train + 1, n_train // 10)):
start_time = time.time()
def test_03_prediction_02(self):
"""Verify that the predictions of the classifier are as expected."""
classifier = PerceptronClassifier({'highly': 1, 'boring': -2})
for instance in self.small_dataset_pred_test_2.instance_list:
prediction = classifier.prediction(instance.feature_counts)
self.assertEqual(prediction, instance.label)
from perceptron import PerceptronClassifier
from arff import Arff
import numpy as np
mat = Arff("standardVoting.arff", label_count=1)
data = mat.data[:, 0:-1]
labels = mat.data[:, -1:]
PClass = PerceptronClassifier(lr=0.1)
PClass.fit(data, labels)
Accuracy = PClass.score(data, labels)
print("Accuray = [{:.5f}]".format(Accuracy))
print("Final Weights =", PClass.get_weights())
def __init__(self, legalLabels, maxIterations):
PerceptronClassifier.__init__(self, legalLabels, maxIterations)
self.weights = util.Counter()
# its settings
if __name__ == "__main__":
# mat = Arff("linsep2nonorigin.arff", label_count=1)
# mat = Arff("data_banknote_authentication.arff", label_count=1)
mat = Arff("votingMissingValuesReplaced.arff", label_count=1)
# mat = Arff("test2.arff", label_count=1)
data = mat.data[:, 0:-1]
labels = mat.data[:, -1].reshape(-1, 1)
X_train, y_train, X_eval, y_eval = split(data,
labels) # split data in 70/30
# PClass = PerceptronClassifier(lr=0.1, shuffle=False, deterministic=10) #initialize perceptron with settings
# PClass = PerceptronClassifier(lr=0.1, shuffle=False)
PClass = PerceptronClassifier(lr=0.1, shuffle=True)
row, col = data.shape # using all zeros initial weight, if not provided, it will do random
initial_weight = np.zeros((col + 1, 1))
PClass.fit(X_train, y_train, initial_weight) # train the perceptron
# PClass.fit(data, labels, initial_weight)
# graph(data[:,0].reshape(-1,1),data[:,1].reshape(-1,1),labels, "linearly separable") # plotting scatter plot
# index = list(range(1, PClass.num_epoch + 1)) # plotting scatter plot with each epoch's error during training
# plt.scatter(index, PClass.epoch_error)
# plt.plot(index, PClass.epoch_error)
# plt.show()
# y = lambda x: -1.3333* x # plotting the line
# y = lambda x: 0.8181 * x
# graph_function(y)
import sys
sys.path.append('../')
from tools import arff
from perceptron import PerceptronClassifier
# arff_file = "data_banknote_authentication.arff"
# arff_file = "dataset_1.arff"
# arff_file = "dataset_2.arff"
arff_file = "voting_data.arff"
mat = arff.Arff(arff_file)
np_mat = mat.data
data = mat[:, :-1]
labels = mat[:, -1].reshape(-1, 1)
#### Make Classifier ####
P2Class = PerceptronClassifier(None, lr=0.1, shuffle=True)
X_train, y_train, X_test, y_test = P2Class._train_test_split(data, labels, 70)
P2Class.fit(X_train, y_train)
misclassifications = P2Class.misclassifications
print('Train split accuracy: {}'.format(P2Class.score(X_train, y_train)))
print('Test split accuracy: {}'.format(P2Class.score(X_test, y_test)))
print("Final Weights =", P2Class.get_weights())
for i in range(len(misclassifications)):
print('epoch: {}, {}'.format(i + 1, misclassifications[i]))
# Accuracy = P2Class.score(data,labels)
# print("Accuray = [{:.2f}]".format(Accuracy))
def __init__(self, legalLabels, maxIterations):
PerceptronClassifier.__init__(self, legalLabels, maxIterations)
self.weights = util.Counter()
elif y[i, 0] == 2:
if out3[i, 0] == 1 and out1[i, 0] == 0 and out2[i, 0] == 0:
correct += 1
elif out3[i, 0] == 1 and net3[i, 0] > net1[i, 0] and net3[
i, 0] > net2[i, 0]:
correct += 1
return correct / row
if __name__ == "__main__":
mat = Arff("iris.arff", label_count=1)
data = mat.data[:, 0:-1]
labels = mat.data[:, -1].reshape(-1, 1)
X_train, y_train, X_eval, y_eval = split(data, labels)
P1 = PerceptronClassifier(lr=0.1, shuffle=True) # train each perceptron
y = process_label(y_train, 0)
row, col = X_train.shape
initial_weight = np.zeros((col + 1, 1))
P1.fit(X_train, y, initial_weight)
print("P1 training accuracy = ", P1.score(X_train, y))
P2 = PerceptronClassifier(lr=0.1, shuffle=True)
y = process_label(y_train, 1)
row, col = X_train.shape
initial_weight = np.zeros((col + 1, 1))
P2.fit(X_train, y, initial_weight)
print("P2 training accuracy = ", P2.score(X_train, y))
P3 = PerceptronClassifier(lr=0.1, shuffle=True)
y = process_label(y_train, 2)
