蘑菇短视频

Skip to main content

MATH 475 Syllabus

Statistical Machine Learning

Revised: October 2021

Course Description

This course blends the algorithmic perspective of machine learning in computer science and the predictive perspective of statistical thinking.  Topics include regression, classification, algorithmic analysis of models, performance metrics and prediction, cross-validation, data transformations, dimension reduction, supervised and unsupervised learning and ensemble methods.
Prerequisites: MATH 270 or MATH 370.
There semester hours.

Student Learning Objectives

By the end of the course, students will be able to:

  • Differentiate between classes of methods and the situations in which they should be applied
  • Compare and contrast the flexibility and interpretability of multiple statistical procedures in the solution of a problem
  • Evaluate and implement multiple statistical learning algorithms to answer a particular question of interest using data
  • Justify the selection of a proposed method through an understanding of the theoretical framework supporting it
  • Apply appropriate computing/programming skills to administer a statistical learning algorithm
  • Communicate the results from a statistical analysis orally and in writing in an appropriate level of detail for an intended audience

Text

Determined by Instructor.

Grading Procedure

Grading procedures and factors influencing course grade are left to the discretion of individual instructors, subject to general university policy.

Attendance Policy

Attendance policy is left to the discretion of individual instructors, subject to general university policy.

Course Outline

Due to the rapidly advancing state of the art in this subject, the instructor has wide latitude in determining the specifics of the course. The following topics should be covered:

  • Regression Techniques
  • Classification Techniques
  • Feature Selection
  • Feature Engineering
  • Exploratory Data Analysis
  • Ensemble Methods
  • Cross-Validation
  • Explainability of models
  • Unsupervised Learning
  • Additional topics may be covered as time and student interest allow