Full Python implementation of a decision tree classifier with pruning. Classifies arXiv abstracts by topic.
Sample usage:
python decision_tree.py -h
python decision_tree.py -i ../Data/train_in.csv -o ../Data/train_out.csv -f 400 -e 1.5 -s 1000
A decision tree will be trained on train_in.csv and evaluated against train_out.csv. It will use 400 words as features and have an entropy threshold of 1.5, it will use 1000 examples to train the tree.
Will produce a file DTmetrics.txt containing the performance metrics of the decision tree classifier.