DecisionTree.py

import pandas as pd
import numpy as np
from sklearn import tree
import pydotplus

# Generate a decision Tree
def createTree(trainingdata):
    data = trainingdata.iloc[:, :-1] #Feature matrix
    labels = trainingdata.iloc[:, -1] # Labels
    trainedTree = tree.DecisionTreeClassifier(criterion="entropy") #Decision tree classifier
    trainedTree.fit(data,labels) # Train the model
    return trainedTree

# Next is to define the function for saving the generated tree diagram

def showtree2pdf(trainedTree,filename):
    dot_data = tree.export_graphviz(trainedTree, out_file=None) # Export the the tree in Graphviz format
    graph = pydotplus.graph_from_dot_data(dot_data)
    graph.write_pdf(filename) #Save the tree diagram to the local machine in PDF format

# define the function for generating Vectorized data

def data2vector(data):
    names = data.columns[:-1]
    for i in names:
        col = pd.Categorical(data[i])
        data[i] = col.codes
    return data

# Now we invoke the function for prediction

data = pd.read_table("./ML/tennis.txt",header=None,sep='\t') # This reads the training data
trainingvec=data2vector(data) # To vectorize data
decisionTree = createTree(trainingvec) # this creates a Decision Tree
showtree2pdf(decisionTree,"tennis.pdf") #to plot the decision tree

# the following predicts a new sample

testVec = [0,0,1,1] #This indicates that the weather is sunny,temperature is low,humidity is high and the wind is strong
print(decisionTree.predict(np.array(testVec).reshape(1,-1)))