Skip to content

harsh1399/CS-6350-Machine-Learning-UoU

Repository files navigation

CS-6350-Machine-Learning-UoU

This is a machine learning library developed by Harsh Mahajan for CS5350/6350 in University of Utah

How to use decision tree?

DecisionTree.py file contains ID3 class. To create a decision tree, create an object of ID3 class with following parameters -

  • Max Depth -> the maximum depth required for decision tree.(Default Value - None)
  • Information gain -> How you want to calculate information gain? (Pass 0 - entropy, 1 - gini value, 2 - majority error)

Methods in decision tree -

  • fit(data,attributeList,depth) data: dataset to train decision tree attributeList: different features(columns) present in our data depth: current depth of decision tree. (Default value = 0)

  • testdataset(testdata) testdata: test dataset to test our model

  • calculateErrors(predictedlabels, targetlabels) predictedlabels: predicted labels from our trained model. targetlabels: actual labels from dataset

To train model-
tree = ID3(None, 0)
tree.fit(data,attributes)

To test dataset-
predicted = tree.testdataset(testdata)

To calculate errors-
tree.calculateErrors(predicted,targetlabels)