Projects

Portland Street Crime Forecasting

• Built a spatial-temporal model that captures the seasonal trend
• Identified the “hot” areas so that prediction accurary index (PAI) is maximized
• Poster available here PDF

Topic Modeling on StackExchange Posts

• Natural language processing on massive data
• Implemented LDA model to cluster posts based on topics
• Report available here PDF

Identify Characters from Google Street View Pictures

• Image processing
• Compared different machine learning algorithms: SVM, Random Forest and Naive Bayes
• Report available here PDF

Identify Robust Economic Growth Determinants

• Built a hierarchical Bayesian model
• Implemented stochastic search variable selection
• Poster available here PDF

Prevalence of Malaria among Children in Gambia

• Built a hierarchical logistic model
• Identified important factors for predicting prevalence of malaria
• Report available here PDF