Newer
Older
import numpy as np
import sys, random
from matplotlib import pyplot as plt
from rf import (
DecisionTree,
Gini, Entropy, MAE, MSE,
Mode, Mean,
CART, ID3, C45,
UsersChoice, Variance, Random, MutualInformation, ANOVA, KendallTau
)
from metric import ConfusionMatrix
from utility import Time
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
def dataShift(dims):
offSet = [0.25, 0.5, 0.25]
diffLen = abs(len(offSet) - dims)
offSet.extend([0] * diffLen)
random.shuffle(offSet)
return offSet[:dims]
def getImurity(impurity: str):
if impurity == 'gini':
return Gini() # Use Gini index as the impurity measure
elif impurity == 'entropy':
return Entropy() # Use Entropy index as the impurity measure
elif impurity == 'mae':
return MAE() # Use MAE index as the impurity measure
elif impurity == 'mse':
return MSE() # Use MSE index as the impurity measure
def getLeaf(leaf: str):
if leaf == 'mode':
return Mode() # Use mode as the leaf function
elif leaf == 'mean':
return Mean() # Use mean as the leaf function
def getSplit(split: str, percentile: int = None):
if split == 'id3':
return ID3(percentile) # Use ID3 algorithm for splitting
elif split == 'c45':
return C45(percentile) # Use C4.5 algorithm for splitting
elif split == 'cart':
return CART(percentile) # Use CART algorithm for splitting
def getFeatureSelection(selection: str, *args):
if selection == 'choice':
return UsersChoice(*args)
elif selection == 'variance':
return Variance(*args)
elif selection == 'random':
return Random(*args)
elif selection == 'mutual':
return MutualInformation(*args)
elif selection == 'anova':
return ANOVA(*args)
elif selection == 'kendall':
return KendallTau(*args)
if __name__ == "__main__":
settings = TreeSettings()
try:
configFile = sys.argv[1]
settings.getConfig(configFile)
settings.setConfig()
except IndexError:
pass
print(settings)
# Create a timer object to measure execution time
timer = Time()
print("Importing data...\n")
timer.start()
data = Data(trainAmount=settings['trainAmount'], evalAmount=settings['validAmount'], dataPath=settings['dataPath'], normalize=settings['normalize'])
data.inputFeatures(*settings['features'])
data.importData(*settings['dataFiles'])
print(data)
timer.record("Importing Data")
# Create and train a decision tree
timer.start()
print('Setting up tree')
tree = DecisionTree(settings['depth'], settings['minSamples'])
tree.setComponent(getImurity(settings['impurity']))
tree.setComponent(getLeaf(settings['leaf']))
tree.setComponent(getSplit(settings['split'],settings['percentile']))
if settings['featSelection'] is not None:
tree.setFeatureSelection(settings['featSelection'], settings['featParameter']) # Use random feature selection
timer.record("Tree setup")
# Train the tree using the training data
timer.start()
print('begin training...')
#tree.train(trainData,trainLabels)
tree.train(data.trainSet.data,data.trainSet.labels.argmax(1))
timer.record("Training")
# Evaluate the tree on the validation data
timer.start()
print('making predictions...')
#prediction = tree.eval(validData)
prediction = tree.eval(data.evalSet.data)
timer.record("Prediction")
# Print the trained decision tree
print(tree)
print()
# Compute confusion matrix to evaluate the performance of the decision tree
confusion = ConfusionMatrix(2)
#confusion.update(prediction, validLabels)
confusion.update(prediction, data.evalSet.labels.argmax(1))
confusion.percentages()
confusion.calcScores()
# Compute confusion matrix to evaluate the performance of the decision tree
print(confusion)
print()
print(timer)