เปรียบเทียบวิธี

	Structural Topic Model	Supervised Text Classification
สาขาวิชา	Political Science	Political Science
ตระกูล	Process / pipeline	Process / pipeline
ปีกำเนิด	2014	2013
ผู้ริเริ่ม	Margaret Roberts, Brandon Stewart & Dustin Tingley	Machine-learning classification tradition (formalized for political text by Grimmer & Stewart; category-proportion variant by Hopkins & King)
ประเภท	Mixed-membership topic model with document-level covariates	Supervised machine-learning classification of documents
แหล่งต้นตำรับ	Roberts, M. E., Stewart, B. M., Tingley, D., Lucas, C., Leder-Luis, J., Gadarian, S. K., Albertson, B., & Rand, D. G. (2014). Structural Topic Models for Open-Ended Survey Responses. American Journal of Political Science, 58(4), 1064–1082.	Grimmer, J., & Stewart, B. M. (2013). Text as Data: The Promise and Pitfalls of Automatic Content Analysis Methods for Political Texts. Political Analysis, 21(3), 267–297.
ชื่อเรียกอื่น	STM, Structural topic modeling, Covariate-aware topic model, Topic model with metadata	Supervised document classification, Text categorization, Automated text coding, Supervised content analysis
ที่เกี่ยวข้อง	5	5
สรุป	The Structural Topic Model (STM) is a text-as-data method that discovers latent themes in a corpus while letting document metadata — party, time, gender, treatment condition — shape those themes. Introduced by Roberts, Stewart, Tingley and colleagues in 2014, it generalizes correlated topic modeling so that topic prevalence (how much a document is about a topic) and topic content (the words used to express a topic) can both depend on covariates. The result is a single model that simultaneously estimates topics and how their use varies across known groups, with uncertainty.	Supervised text classification trains a statistical model on documents that humans have hand-labeled, then uses it to assign categories — topic, tone, position, relevance — to the much larger set of unlabeled documents. Unlike dictionary methods, which apply a fixed word list, a supervised classifier learns from examples which textual features predict each category, so it can capture context-dependent and non-obvious cues. Grimmer and Stewart present it as a core text-as-data workflow, and a key insight is that for many political-science questions the goal is not perfect document-by-document labels but accurate estimates of category proportions across a corpus.