The SAGE Handbook of Research Methods in Political Science and International Relations offers a comprehensive overview of research processes in social science - from the ideation and design of research projects, through the construction of theoretical arguments, to conceptualization, measurement, and data collection, and quantitative and qualitative empirical analysis - exposited through 65 major new contributions from leading international methodologists. Each chapter surveys, builds upon, and extends the modern state of the art in its area. Following through its six-part organization, undergraduate and graduate students, researchers and practicing academics will be guided through the design, methods, and analysis of issues in Political Science and International Relations: Part One: Formulating Good Research Questions and Designing Good Research Projects; Part Two: Methods of Theoretical Argumentation; Part Three: Conceptualization and Measurement; Part Four: Large-Scale Data Collection and Representation Methods; Part Five: Quantitative-Empirical Methods; Part Six: Qualitative and Mixed Methods.
Chapter 56: Machine Learning in Political Science: Supervised Learning Models
Machine Learning in Political Science: Supervised Learning Models
The set of theoretical and computational approaches used under the rubric of ‘machine learning’ (ML) is so diverse that it is easy to think of the field as a kind of catch-all, interdisciplinary exercise at the intersection between statistics, computer science and other affiliated disciplines. Such a perspective, however, would obscure the fact that these seemingly disparate approaches share a common goal: to improve a computer's performance on a given (typically predictive) task by identifying empirical relationships, patterns, and trends in data that rely on minimal distributional and functional form assumptions on the part of analysts (Hastie et al., 2009; Jordan and Mitchell, 2015). This goal, which is what learning is typically taken to ...