Comparison of LSI algorithms without and with pre-processing: using text document based search

Searching of document/text is the most important need of each student or computer user. Searching through particular index or term is the old fashion, now a day's user want to search documents according to some phrase, query or requirement i.e. extraction of meaningful information from large collection according to some textual query. Different methods such as iterative residual rescaling (IRR), term frequency (TF), inverse document frequency (IDF), multi words are using to handle such issues. Latent semantic indexing (LSI) is an important method for current literature of information retrieval. LSI can find similar documents on particular textual phrase. Here author has implemented two algorithms (without and with pre-processing) of LSI for text documents. As a result, both algorithms can obtain the similar results but their processing time will be different.