Fall 2018 Projects

Anonymous Leading Semiconductor Manufacturing Equipment Maker: Laser Failure Prediction Using Time-series Modeling

Team members: Xingyu Wang, Charlene Lau, Xiang Li, Ansheng Xu

Students developed predictive models to model laser failure using measured laser parameters. The team took a novel approach to wrangle the time-series data they were provided with. The models they created predicted laser failure date and probability of laser failure at any point in time. They also compared the performance of machine learning regression algorithms to that of classic curve fitting techniques.

Harris RF: Supplier Risk Evaluation Using Clustering Models

Team members: Adrian Eldridge, Gazi Naven, Peter Chen, Trevor Whitestone

Students applied clustering models to group suppliers based on risk-level, and identified characteristics of supply risk. Team members employed feature engineering, dimensionality reduction, and factor analysis, and constructed clustering models for risk analysis.

Martino Flynn: Predict Insurance Policy Lapses For Purchases From Direct TV Advertising

Team members: Xiling Shen, Yang Wang, Yifan Jiang, Yuanqi Zhu

Students utilized machine learning algorithms to predict insurance policy lapses for purchases from Direct TV advertising. They applied exploratory analysis, feature selection, model construction, model validation, and model improvement to the data they were given, and provided Martino Flynn with useful business insights.

Paychex: Predict Quality of Service (QoS) Issues Using Machine Learning

Team members: Zhuoyou Wang, Zhenkai Liu, Xuexun Xiao, Benjamin Chen

Students employed machine learning algorithms to predict Quality of Service (QoS) issues for clients, using the sponsor’s IT severs to identify the root causes of problems. The team pre-processed data and performed exploratory analysis for insights. They then constructed predictive models and evaluated model performance.

RTS: Evaluate Bus Operator Absenteeism Patterns Using Machine Learning Techniques

Team members: Chester Szeen, Kishore Thummuluru, SivaDurga Thotapalli, Sofia Salen

Students applied data visualization tools and machine learning algorithms to data to identify trends in bus operator absenteeism. The team used data visualization to identify absenteeism trends over time, and applied classifications to data to predict absenteeism, utilizing geographical and work scheduling features.

Xerox: Predict Upsell Opportunities For Managing Printer Fleets

Team members: Haiting Zhu, Xuejian Ye, Yujian Wu, Yunrou Gong

Students pre-processed a large data set and developed models that can forecast opportunities for Xerox to manage larger printer fleets for current clients. Team members applied data visualization tools to identify clients with potential opportunities for Xerox. Based on revenue increase analyses, students could provide executable insights about volume-of-growth opportunities for each client.