Vector Model Based Information Retrieval System With Word Embedding Transformation

Vector based information retrieval system has been one of the trending methods in Natural Language Processing. The embeddings vector generated from a document helps in identifying most relevant document related to the query. There is various approach were embedding vectors can be generated and some of them which have implemented are Word2vec, Glove2vec and Sentence BERT. For information retrieval system also used word embedding transformation like PCA and Factor Analysis to improvise the model's performance. Most of information retrieval system involves getting query from the user, preprocessing of the query and generating most relevant information to the query. Results obtained by post processing methods such as PCA and Factor Analysis shows a comparatively better results with an increase of 2–3% of Mean average precision.