Udo Nweke

Master's

Computer Science

Old Dominion University

Research Assistant

Projects • Analyzed Customer Purchase Behavior based on Purchase History for Various Products: Developed a model that understands customer purchase pattern against various products of different categories utilizing customer purchase history and demographics • Credit Card Fraud Detection: Utilized different classification algorithm (Naïve Bayes, Random Forest, Logistic Regression, and Support Vector Machine) to identify fraud transaction on the given dataset. Random Forest Model gave a higher degree of comprehensiveness with better performance. • Movie Recommendation System: Reviewed a paper that provided a method for processing implicit feedback and implemented the algorithm by building a recommendation system for television shows based on analysis done on watching habits for anonymized users • Kaggle Zillow Home Value Prediction Competition: Developed a Light GBM algorithm using decision tree based model to process over a million dataset with 58 features and full list of real estate properties in three different cities in United States (Los Angeles, Orange and Ventura) • Immigration Data Visualization: Used storytelling to visualize the effect of immigration policies and events in the United States within a window of time.