/Web-Mining-and-Recommender-Systems

My work on UCSD CSE 258 Web Mining and Recommender Systems

Primary LanguageJupyter NotebookMIT LicenseMIT

Web-Mining-and-Recommender-Systems

My work on UCSD CSE 258 Web Mining and Recommender Systems

Homework 1 Regression and Classification Tasks

Analyzed data from 50,000 beer reviews. Then trained a simple predictor with a single binary feature indicating whether a beer is an `American IPA', Evaluate result using MSE.

Homework 2 Classifier Evaluation and Dimensionality Reduction

Implemented weighted 2-class logistic regression by gradient ascent to solve the issue of class imbalance (i.e., an uneven number of the datapoints have a positive label). Then applied PCA dimensionality reduction to try to find the principal components of the word features.