Book Recommendation System

Course Project for Columbia University : E4571 Personalization - Theory & Application
Team members: Deepak Maran, Kewei Liu, Rakshita Nagalla, Xiaohui Guo

For detailed explanation of the approach and discussion of the results, refer to this ipython notebook

Contents:

BX-users.csv: Data file with information about the users
train.csv: Data used for validation and training
test.csv: Data used for testing image: Folder containing all the plots

Data Cleaning and Splitting.py: Cleans the BX-Book-ratings.csv file and splits into train and test files
Hyperparameter Tunning_1.ipynb: Cross-validation code for tuning latent factors and learning rate hyperparameters for SGD
Hyperparameter Tunning_2.ipynb: Cross-validation code for tuning regularization term hyperparameter for SGD
MAP_inference_Coordinate_Ascent_Algorithm.py: Code for implementation of Probabilistic Matrix Factorization MAP inference coordinate ascent algorithm.
kNN_v5.py: Cross-validation code for tuning hyperparameters for kNN
evaluation.py: Training and test set accuracy and coverage for SGD, kNN and baseline models. Precision-recall curves also plotted
SampleSizeTime.py: Plots running time as a function of sample size
sampleSizeAccuracy.py: Plots accuracy as a function of sample size

Software_Package_Requirements.txt: Software requirements to execute the code