Causal Latent Semantic Analysis (cLSA): An Illustration

Abstract

Latent semantic analysis (LSA), a mathematical and statistical technique, is used to uncover latent semantic structure within a text corpus. It is a methodology that can extract the contextual-usage meaning of words and obtain approximate estimates of meaning similarities among words and text passages. While LSA has a plethora of applications such as natural language processing and library indexing, it lacks the ability to validate models that possess interrelations and/or causal relationships between constructs. The objective of this study is to develop a modified latent semantic analysis called the causal latent semantic analysis (cLSA) that can be used both to uncover the latent semantic factors and to establish causal relationships among these factors. The cLSA methodology illustrated in this study will provide academicians with a new approach to test causal models based on quantitative analysis of the textual data. The managerial implication of this study is that managers can get an aggregated understanding of their business models because the cLSA methodology provides a validation of them based on anecdotal evidence.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 28 citations exist upstream.