A Technique to Improve Difficult Keyword Queries Over Database

Estimating query performance is the job of predicting the excellence of results returned to examine in response of a query. queries on databases provide easy access to data, but often it goes through from low ranking quality. It is defined to get queries with low ranked result, quality to improve the user satisfaction. Post-retrieval predictors analyze the o of top-retrieved documents. This paper introduced a new technique to get high-performance named as NASA, this method is based on k-Nearest Neighbor (k-NN) search on the top-k results of the corrupted version of database. k-NN handles complex functions during the execution and it improve the loss of information. Simultaneously it helps to reduce the execution time.