Latent Semantic Analysis(LSA) method and theory for automatic indexing and implicit higher-order structure of connections term-by-document using singular-value decomposition technique for finding relevant documents on basis of terms found in queries. The main idea is that document is set of words, which have only one idea. The order of words is ignored, only take a note how many times particular word is represented in one document. Through words, which is terms LSA find other documents which have same words – same meaning.
Latent Semantic Analysis, term-by-document matrix, Singular Value Decomposition, Stemming, Dimension reduction, machine learning, document classification, unsupervised machine learning
