Modification of Keyword Selection Process to Get Least List with Weighted Keywords by using Essence of both ‘Baxendale’ and ‘Swanson’ Experiment

The ‘keyword-based search system’ by its name implies that the document is presented with the correct keyword. Keyword-content relations need to be thoroughly analyzed in order to select a document's keywords. In this case, the current method used, even after the use of ontology, has not been adequate. That's why many keywords are being selected for a particular document. And users are bounded to use all those keywords in search of a single document, resulting in a decrease in the precision of retrieval day by day. So in this case it is not possible to get high precision and to reduce quantity of keywords without selecting high quality keywords. The purpose of this paper is to examine how informative a particular keyword is, and to select the highest and most informative keyword and to exclude the least informative keyword through which, ‘keyword-based search system’ can get fewer but higher quality or weighted keywords as a helping hand to solve ‘High Recall and Low Precision’.