Latent Semantic Analysis for Text-Based Research

Latent Semantic Analysis(LSA) is a theory and method for extracting and representing the contextual-usage meaning of words by statistical computations applied to a large corpus of text.It is proved to be an efficient method in text-based research from three aspects: Matching summaries to the texts read,essays grading and the measurement of textual coherence,among which the last one plays the key role.LSA can measure the amount of semantic overlap between adjoining sections of text to calculate coherence.