Precision improvement of ranked text retrieval using keyword weighting on paragraph vector

The needs to find and retrieve relevant material in order to satisfy information needs from a collection of documents increase rapidly. Some institutions need to find and retrieve documents from stored previous documents to solve problems. An approach method can be applied to retrieve relevant documents from a set of document is Paragraph Vector. This method calculates documents similarity based on two unique vectors of a document and uses the similarity to rank the relevant retrieved documents. For increase document ranking, term weighting can be combined within Paragraph Vector to make meaningful words such as keywords became a higher priority in process of retrieval. The proposed method combines weighting of keywords with Paragraph Vector to improve similarity between documents based on the frequency of keyword occurrence.