STAT 535: Statistical Machine Learning Syllabus
Uploaded by
zwu363STAT 535: Statistical Machine Learning Syllabus
Uploaded by
zwu363Common questions
Powered by AICausal inference in STAT 535 likely integrates with machine learning to help distinguish correlation from causation. This is fundamental in many applications, such as policy making or medical studies, where understanding causal relationships can directly impact decisions and outcomes. Thus, the course emphasizes equipping students with tools to infer causality amidst predictive modeling .
Students are required to have a solid understanding of calculus, linear algebra, probability, and mathematical statistics. Additionally, a good sense of mathematical analysis is necessary. These prerequisites ensure that students can grasp the mathematical concepts that underlie statistical machine learning techniques .
Programming is integral to the homework assignments in STAT 535, requiring students to include code in their solutions. However, students are restricted from using certain advanced packages, specifically e1071 and kernlab in R, unless approved. Allowed packages are for basic operations such as linear algebra (e.g., numpy in Python) and k-nearest neighbor search (RANN in R).
The University of Washington, including the School of Public Health, enforces strict academic integrity policies. Violations such as plagiarism and cheating are serious and will be dealt with according to university guidelines. Maintaining academic conduct is essential, and any misconduct is subject to disciplinary actions as per the UW Student Conduct Code .
The evaluation of student performance in STAT 535 consists of three key components: homework assignments (50%), a final project report (20%), and a final exam (30%). The homework is crucial as it constitutes half of the overall grade, highlighting its importance in reinforcing course material .
Nonparametric density estimation and regression, as taught in STAT 535, can be applied to real-world scenarios like financial risk modeling, where data distributions aren't assumed to follow a specific parametric form. In medical diagnosis, nonparametric regression helps in predicting patient outcomes without assuming linear relationships, thereby accommodating more complex patterns in the data .
Accommodations in STAT 535 are guided by policies for inclusivity and accessibility. Students needing accommodations should activate them via myDRS. For religious accommodations, requests must be made within the first two weeks of the course using the designated form. The aim is to address diverse needs, including those related to disabilities and religious practices .
Focusing on both nonparametric classification and clustering offers a comprehensive understanding of methods that do not rely on parametric assumptions of data distributions. This is crucial in scenarios where data does not fit traditional models, allowing flexibility and robustness in analyzing complex datasets. These topics illustrate the ability to uncover patterns and groupings in diverse data sets, enhancing the applicability of statistical learning techniques .
The topic of missing data is explicitly included within the course content of STAT 535. Although the syllabus does not detail the methods taught, the inclusion suggests a focus on understanding and applying statistical approaches to dealing with incomplete datasets, a common challenge in data analysis .
The final project involves summarizing one or several published papers related to the course topics. Groups of 1-3 students should submit a report in PDF format using the NeurIPS template. The report, which must be 4-6 pages long with up to 2 additional pages for the appendix, should clearly convey the results of the papers and any new insights if applicable, which could result in extra points .