Teaching

University of Massachusetts Amherst, School of Public Health and Health Sciences

BIOSTATS 743: Analysis of Categorical Data in Public Health, Teaching Assistant

• Semester: 2024Fall
• Course Description:
  • This course provides an overview of statistical methods for analyzing data where the outcome variable is categorical or discrete. The course will emphasize the theoretical underpinnings of the methods as well as an applied understanding of the computation and interpretation, both of which are necessary to succeed with real data analysis. We will cover inference for binomial and multinomial variables with contingency tables, generalized linear models, logistic regression for binary responses, logit models for multiple response categories, log-linear models, some statistical machine learning approaches, inference for matched-pairs, and correlated/clustered data. Examples will be taken from public health and biomedical research.

University of Illinois Urbana-Champaign, Department of Statistics

STAT 432: Basics of Statistical Learning, Course Assistant

• Semester: 2023Fall
• Course Description:
  • This course covers a wide range of topics in machine learning, including both supervised and unsupervised learning techniques. Key areas of focus include:
    • Supervised Learning:
      • Linear models and penalization methods to prevent overfitting.
      • Discriminant analysis, Naive Bayes, logistic regression for classification.
      • K nearest neighbor, classification and regression trees, random forest, kernel regression for decision-making processes.
      • Support vector machines for high-dimensional learning.
    • Unsupervised Learning:
      • PCA, K-mean and hierarchical clustering, self-organizing maps, spectral clustering for data structure discovery.
    • Concepts:
      • Understanding the bias-variance trade-off and its implications on model accuracy.
      • Techniques for variable selection and model optimization.
      • Methods like cross-validation and bootstrap for model evaluation and enhancement.
    • Modeling Problems:
      • Applications in personalized medicine and imaging data analysis.