Course Projects

biostatistic method

• Mastered the basic biostatistics methods.
• Correlation analysis, analysis of variance (ANOVA), regression analysis(SLR, MLR, Logistic regression), etc.
• Parametric test like 2-sample t-test and one-way anova or nonparametric test like sign test, Wilcoxon signed rank test, rank sum test and Fisher’s Exact test.
• Here is a group project we use K-M curve, cox ph model and logistic regression to predict breast cancer patient’s survival rate.

data science

• Familiarized with basic R programming skills.
• Learned to use R for interaction with other applications, such as Shiny and web scraping.
• Here is a group presentation in class analyzing the corruption index. I use shiny and make a website.
• I studied basic machine learning knowledge.
• I studied many model training methods: Lasso, Ridge, Elastic Net, K-Nearest Neighbors (KNN), Principal Components Regression (PCR), Generalized Additive Model (GAM), Multivariate Adaptive Regression Splines (MARS)
• Here is a project exploring the factors associated with recovery time for COVID-19 patients.
• Classfication tree, Probabilistic Models (Linear Discriminant Analysis, Quadratic Discriminant Analysis and Naive Bayes) and Kernal-based model (Support Vector Machines with Linear Kernel and Support Vector Machines with Radial Kernel) are also included.
• Here is a project exploring the factors affecting the severity of COVID-19.

advanced statistical computing

• I studied Monte-Carlo-methods, MC integration, optimization methods(Newton’s method, EM algorithm), resampling methods(bootstrap) and Markov Chain Monte Carlo(MCMC).
• Here is a project comparing bootstrap methods for propensity score matching.
• Here is a project exploring single-cell gene expressions and clustering.

longitudinal analysis

• I learned the transition from linear regression with independent observations to longitudinal data analysis with correlated observations.
• I studied the difference between marginal model (GEE/gls) and the GLMM/GLM model (Max likelihood).
• I studied how to deal with missing data.
• Here is a project exploring wihch model (GLMM or marginal model) is better when analyzing women’s contraception.

survival analysis

• I learned non-parametric methods in survival analysis, such as the Kaplan-Meier (K-M) curve, semi-parametric methods like the Cox Proportional Hazards (Cox PH) model, and parametric methods, assuming Y follows distributions such as Weibull, Exponential, Log-Logistic, or General Gamma.
• Here is a project exploring the factors affecting the survival risk for Breast Cancer.