11-865 Tracking Political Sentiments Using Machine Learning
12 units, Carnegie Mellon University
Online social media provides a rich source of detailed data reflecting the evolution of political sentiments over time, and in response to various news events. This seminar course studies a diverse set of recent papers in the intersection of machine learning, natural language processing and political science with an aim to pose nontrivial research questions concerning US politics and devise ML and NLP framework for answering them. A key component of the course is a semesterlong research project with the view towards a peerreviewed publication. The course provides a large text data set relevant to US politics and the goal is for each student to formulate, explore and understand a focused research question through the lens of this data set. By the end of the course, apart from acquiring handson experience in realizing the synergy between large scale data, creative research questions and effective NLP solutions, we all hope to have an improved understanding of why we are in what we are in. The course will provide some useful data sets and a list of potential project ideas for its primary focus the semesterlong project. Students will work in small teams 12 members on course projects. The course engineer will provide useful scripts and code to efficiently process data. The course requires familiarity with machine learning at a level of being able to complete a substantial project. Any of the following courses can serve as a prerequisite 10601, 10701, 10605, 10805, 11685, 11785, 95845. Interested students without this requirement should contact the instructor Ashique KhudaBukhsh, akhudabucs.cmu.edu to check if they have the required background.
Prerequisites: none listed
Courseweave needs JavaScript to show sections, workloads and planning tools.