Applying Machine Learning Algorithms using R, Python, and RapidMiner.
This project involves predicting clicks for on-line advertisements. The training data consists of data for 9 days from October 21, 2014 to October 29, 2014. The criterion used to evaluate the performance is log-loss. We used lasso regression, K-nearest Neighbor Classification, Decision Tree Classification and ensemble method to predict the probability of clicks. Contributors: Lancy Mao, Athena Li (Teammate)
This project involves predicting spam emails. There are two parts: 1: used feature selection, naive bayes model, logistic regression, random forest model, and ensemble method; 2: used regularized models with penalties--lasso and ridge regression. Contributors: Lancy Mao, Athena Li (Teammate), George Easton (Professor)
This project involves predicting sales ranking of ebooks on Kindle based on their features. Contributors: Lancy Mao, Athena Li (Teammate)
This exercise involves predicting breast cancer. Performed a predictive modeling analysis on this same dataset using the decision tree, the k-NN technique and the logistic regression technique. Contributors: Lancy Mao, Vilma Todri (Professor)
This exercise involves predicting customer spending. Used Rapidminer to build numeric prediction models that predict Spending based on the other available customer information. Used linear regression, k-NN, and regression tree techniques. Contributors: Lancy Mao, Vilma Todri (Professor)