CSC466FinalProject

Aditi Gajjar, Anagha Sikha, Othilia Norell, Soren Paetau, Nicholas Tan \ agajjar@calpoly.edu / arsikha@calpoly.edu \ onorell@calpoly.edu / spaetau@calpoly.edu / nktan@calpoly.edu ​

HOW TO RUN CODE

Q1: (Random Forest on Feature Importance)

• run python3 randomForest.py To run code on different subsets of attributes (all, without best attributes, and without worst attributes) by updating line 34 in randomForest.py – instructions in randomForest.py

Q2 (Apriori Rules): All code is contained within a jupyter notebook

Q3 (Clustering):

• For KMeans, run python3 kmeans.py spotify_songs.csv [-p (kmeans_plus)] [-m (manhattan dist)] [-n (normalize data)] [-t (testing)]
• For DBScan, run python3 dbscan.py spotify_songs.csv [-m (manhattan dist)]

Q4 (Decision Tree vs Random Forest):

• To run C4.5 run python3 InduceC45.py spotify_new_train.csv in the terminal
• To run classify on the decision tree from InduceC45.py run on the test dataset: python3 classify.py spotify_new_test.csv spotify_new_train.json
• To run Random Forests run python3 randomForest.py spotify_songs_new.csv (example: python3 randomForest.py spotify_songs_new.csv 9 600 50)
• Results of the random forests classification are found in spotify_songs_new_rf_all.txt for all 9 attributes
• From hypertuning results of the random forests classification are found in spotify_songs_new_rf.txt for 4 attributes

Q5 (Collaborative Filtering):

-Reference readme in CollabFiltering -use -sp with testcases given