Computer-Assisted Text Analysis for Social Science: Topic Models and Beyond.

摘要: Topic models are a family of statistical-based algorithms to summarize, explore and index large collections text documents. After decade research led by computer scientists, topic have spread social science as new generation data-driven scientists searched for tools unstructured text. Recently, contributed model literature with developments in causal inference handling the problem multi-modality. In this paper, I provide review on evolution modeling including extensions document covariates, methods evaluation interpretation, advances interactive visualizations along each aspect's relevance application research.