# shannonEnt = Trees.calcShannonEnt(myDat) # print shannonEnt # retDataSet = Trees.splitDataSet(myDat, 0, 1) # print retDataSet # bestFeature = Trees.chooseBestFeatureToSplit(myDat) # print bestFeature # tree = Trees.createTree(myDat, labels) # print tree ''' # TreePlotter ''' # TreePlotter.createPlot() tree = TreePlotter.retrieveTree(0) # numLeafs = TreePlotter.getNumLeafs(tree) # depth = TreePlotter.getTreeDepth(tree) # print "leafs nums: %d, depth: %d" % (numLeafs, depth) TreePlotter.createPlot(tree) ''' # Classify ''' # myDat, labels = Trees.createDataSet() # myTree = TreePlotter.retrieveTree(0)
0: 'lef node',
1: {
'level 2': {
0: 'leaf node',
1: 'leaf node'
}
},
2: {
'lead 2': {
0: 'leaf node',
1: 'leaf node'
}
}
}
}
tp.createPlot(myTree)
# 图与网格结构的可视化
data = np.mat([[0.1, 0.1], [0.9, 0.5], [0.3, 0.6], [0.7, 0.2], [0.1, 0.7],
[0.5, 0.1]])
m, n = np.shape(data)
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
ax.scatter(data.T[0].tolist(), data.T[1].tolist(), color='blue', marker='o')
for point in data.tolist():
plt.annotate("(" + str(point[0]) + "," + str(point[1]) + ")",
xy=(point[0], point[1]))
xList = []
yList = []
for px, py in zip(data.T.tolist()[0], data.T.tolist()[1]):
# shannonEnt = Trees.calcShannonEnt(myDat) # print shannonEnt # retDataSet = Trees.splitDataSet(myDat, 0, 1) # print retDataSet # bestFeature = Trees.chooseBestFeatureToSplit(myDat) # print bestFeature # tree = Trees.createTree(myDat, labels) # print tree ''' # TreePlotter ''' # TreePlotter.createPlot() tree = TreePlotter.retrieveTree(0) # numLeafs = TreePlotter.getNumLeafs(tree) # depth = TreePlotter.getTreeDepth(tree) # print "leafs nums: %d, depth: %d" % (numLeafs, depth) TreePlotter.createPlot(tree) ''' # Classify ''' # myDat, labels = Trees.createDataSet() # myTree = TreePlotter.retrieveTree(0)
:param filename: 存储的文件
:return:
"""
import pickle
with open(filename, 'w') as fw:
pickle.dumps(input_tree, fw)
def grab_tree(filename):
"""
从文件中读取决策树
:param filename: 要读取的文件
:return: 决策树
"""
import pickle
with open(filename) as fr:
return pickle.load(fr)
if __name__ == '__main__':
with open('lenses.txt') as fr:
lenses = [inst.strip().split('\t') for inst in fr.readlines()]
# 标签列表
lensesLabels = ['age', 'prescript', 'astigmatic', 'tearRate']
# 创建决策树
lensesTree = create_tree(lenses, lensesLabels)
# 打印树
print(lensesTree)
# 显示树形图
TreePlotter.create_plot(lensesTree)
pickle.dump(obj, file)
file.close()
def readDump(path):
file = open(path, "rb")
data = pickle.load(file)
file.close()
return data
dtree = ID3Tree()
if not os.path.exists(ID3SavePath): # 不是文件夹
print("生成数据")
dtree.dataSet = loadDataSet(ID3LoadPath)
dtree.labels = ["age", "revenue", "student", "credit"]
# dtree.dataSet = dataSet
# dtree.labels = labels
print("训练数据")
dtree.train()
print("持久化数据")
saveDump(ID3SavePath, dtree.getDumpData())
else:
print("读取持久化")
dtree.loadDumpData(readDump(ID3SavePath))
print("正在生成树")
tp.createPlot(dtree.tree)
print("预测结果为:", dtree.predict(dtree.tree, [0, 0, 0, 0]))
2. Prepare: Parse tab-delimited lines.
3. Analyze: Quickly review data visually to make sure it was parsed properly. The final
tree will be plotted with createPlot().
4. Train: Use createTree() from section 3.1.
5. Test: Write a function to descend the tree for a given instance.
6. Use: Persist the tree data structure so it can be recalled without building the
tree; then use it in any application.
'''
from numpy import *
import LoadData as ld
import DicisionTree as dts
import TreePlotter as tplt
dataSet, labels = ld.createDataSet('lenses.txt')
#testLabels = zeros(len(labels),1)#也不能采用
#testLabels = labels# 不能采用
lensesTree = dts.createTree(dataSet, labels)
print lensesTree
tplt.createPlot(lensesTree)
print labels
#利用训练数据做测试数据
dataSet, testLabels = ld.createDataSet('lenses.txt')
errorCount = 0
numTestVec = 0
for testVec in dataSet:
numTestVec += 1.0
classLabel = dts.classify(lensesTree, testLabels, testVec)
if classLabel != testVec[-1]:
errorCount += 1
errorRate = (float(errorCount) / numTestVec)
print "the error rate of this test is: %f" % errorRate
[u'T', u'D', u'P', 'yes'], [u'D', u'D', u'B', 'yes'],
[u'D', u'N', u'B', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'D', u'A', u'D', 'yes'],
[u'T', u'D', u'P', 'yes'], [u'T', u'D', u'B', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'T', u'N', u'B', 'yes'], [u'D', u'A', u'D', 'yes'],
[u'T', u'D', u'D', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'T', u'D', u'D', 'yes'], [u'D', u'A', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'A', u'D', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'T', u'D', u'B', 'yes'],
[u'T', u'D', u'P', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'T', u'D', u'D', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'T', u'D', u'B', 'yes'],
[u'T', u'P', u'B', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'A', u'P', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'D', u'D', u'P', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'T', u'D', u'D', 'yes'], [u'T', u'D', u'D', 'yes'],
[u'T', u'D', u'D', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'T', u'N', u'B', 'yes'], [u'D', u'N', u'B', 'yes'],
[u'T', u'D', u'D', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'D', u'D', 'yes'], [u'D', u'D', u'D', 'yes'],
[u'D', u'A', u'D', 'yes'], [u'T', u'N', u'B', 'yes']
] #[[1,1,'yes'],[1,1,'yes'],[1,0,'no'],[0,1,'no'],[0,1,'no']]
labels = ['surfacing', 'flippers', 'ff']
myTree = CreatTree(dataset, labels)
print myTree
TreePlotter.CreatPlot(myTree)
import operator import os import Trees myData,labels=Trees.createDataSet() print(myData) print(labels) print(Trees.calcShannonEnt(myData)) myData[0][-1]='maybe' print(myData) print(labels) print(Trees.calcShannonEnt(myData)) myDat,labels=Trees.createDataSet() print(myDat) print(Trees.splitDataSet(myDat,0,1)) print(Trees.splitDataSet(myData,0,0)) print(Trees.chooseBestFeatureToSplit(myDat)) myTree=Trees.createTree(myDat,labels) print(myTree) import TreePlotter TreePlotter.createPlot()
fw.close()
def grabTree(filename):
import pickle
fr = open(filename,"rb")
return pickle.load(fr)
dataSet,labels = createDataSet()
# shannonEnt = calcShannoEnt(dataSet)
# print(shannonEnt)
# retDataSet = splitDataSet(dataSet,0,1) # 取出第0个特征为1的数据,并去掉该特征
# print(dataSet)
# print(retDataSet)
# bestFeature = chooseBestFeatureToSplit(dataSet)
# print(bestFeature)
# myTree = createTree(dataSet,labels)
# print(myTree)
# 使用算法
fr = open('/Users/lixiwei-mac/Documents/IdeaProjects/MachineLearningInAction/DecisionTree/lenses.txt')
lenses = [inst.strip().split('\t') for inst in fr.readlines()]
lensesLabels = ['age','prescript','astigmatic','tearRage']
lensesTree = createTree(lenses,lensesLabels)
print(lensesTree)
treePlotter.createPlot(lensesTree)
maxMJJ=610.0
minMVV=1000.0
maxMVV=7000.0
binsMJJ=290
binsMVV=160
cuts['acceptance']= "(jj_LV_mass>{minMVV}&&jj_LV_mass<{maxMVV}&&jj_l1_softDrop_mass>{minMJJ}&&jj_l1_softDrop_mass<{maxMJJ})".format(minMVV=minMVV,maxMVV=maxMVV,minMJJ=minMJJ,maxMJJ=maxMJJ)
cuts['acceptanceGEN']='(jj_l1_gen_softDrop_mass>0&&jj_gen_partialMass>0)'
cuts['nonres'] = '1'
cut='*'.join([cuts['common'],cuts['nonres'], '(jj_l1_softDrop_mass>30&&jj_l1_softDrop_mass<610)','(jj_LV_mass>1000&&jj_LV_mass<7000)',cuts['HP'],'(jj_l1_gen_softDrop_mass>0&&jj_gen_partialMass>0)'])
dataPlotters=[]
dataPlottersNW=[]
dataPlotters.append(TreePlotter(fromsample+'.root','tree'))
dataPlotters[-1].setupFromFile(fromsample+'.pck')
dataPlotters[-1].addCorrectionFactor('xsec','tree')
dataPlotters[-1].addCorrectionFactor('genWeight','tree')
dataPlotters[-1].addCorrectionFactor('puWeight','tree')
data=MergedPlotter(dataPlotters)
sampleHisto=dataPlotters[0].drawTH2("jj_l1_gen_softDrop_mass:jj_LV_mass",cut,"1",binsMVV,minMVV,maxMVV,binsMJJ,minMJJ,maxMJJ,"M_{qV} mass","GeV","Softdrop mass","GeV","COLZ" )
sampleHisto1Dmjet=dataPlotters[0].drawTH1('jj_l1_softDrop_mass',cut,"1",binsMJJ,minMJJ,maxMJJ)
sampleHisto1Dmjet.Scale(1/sampleHisto1Dmjet.Integral())
print sampleHisto
# 获取信息增益最大的特征及其增益
highest_gain_feature, highest_gain = get_feature_with_highest_Gain(
data_set)
# 增益小于ε，单一节点，返回实例数最大的类
if highest_gain < eps:
return get_most_common_class(data_set)
# 构建树
decision_tree_dict = {highest_gain_feature: {}}
# 对每个最高增益特征的取值进行分割数据集，并进行递归调用生成树
feature_values = set(data_set[highest_gain_feature])
for one_value in feature_values:
# 分割D
divided_data_set = data_set[data_set[highest_gain_feature] ==
one_value]
# 去除列，A = A - {Ak}i
divided_data_set = divided_data_set.drop(labels=highest_gain_feature,
axis=1)
# 生成子树
decision_tree_dict[highest_gain_feature][
one_value] = generate_decision_tree(divided_data_set, eps)
return decision_tree_dict
if __name__ == '__main__':
data_set = init_data('resources/lenses.txt')
decision_tree = generate_decision_tree(data_set, eps=0.0001)
print(decision_tree)
TreePlotter.createPlot(decision_tree)
# print(data_set[(data_set['tearRate'] == 'normal') & (data_set['astigmatic'] == 'yes') & (data_set['prescript'] == 'myope')])
# print(data_set[data_set['tearRate'] == 'reduced'])
def lensesStudy(filepath):
fr = open(filepath)
lenses = [inst.strip().split("\t") for inst in fr.readlines()]
lensesLabels = ['age', 'prescript', 'astigmatic', 'tearRate']
lensesTree = createTree(lenses, lensesLabels)
TreePlotter.createPlot(lensesTree)
def lensesStudy(filepath):
fr = open(filepath)
lenses = [inst.strip().split("\t") for inst in fr.readlines()]
lensesLabels = ['age', 'prescript', 'astigmatic', 'tearRate']
lensesTree = createTree(lenses, lensesLabels)
TreePlotter.createPlot(lensesTree)
