Decision Tree, Bagging Trees and Random Forests Implementation from scratch in Python.
Execute the following scripts to see the results:
python preprocess-assg4.py to preprocess and split the data into training and test set.
python trees.py trainingSet.csv testSet.csv 1 to run decision tree model.
python trees.py trainingSet.csv testSet.csv 2 to run bagging tree model.
python trees.py trainingSet.csv testSet.csv 1 to run random forest model with sampled sqrt(p) attributes.
python cv_depth.py to check the performance of models with varying depth limit of the trees.
python cv_frac.py to plot the learning curves of the three models.
python cv_numtrees.py to compare the performance of models with varying number of trees in the ensemble.
python neural_net.py trainingSet.csv testSet.csv to check the performance of dataset with neural networks.
Results for the above models and scripts are available in hw4.pdf