Cluster Based Prediction of Keyword Query Over Databases

In this paper, We using the cluster-based prediction, to predict the keywords over database in a efficient way. By using the cluster-based prediction, efficiency of searching query is improved and time complexity is reduced. In this paper, we proposed Text preprocessing, MVS matrix, k-means clustering and character shuffle preprocessing searching algorithm in order to improve efficiency. In text preprocessing, we eliminates all the tags and find the relative frequencies of each document then weights is calculated. By using MVS matrix, similarities of each document is calculated then formed into matrix. Based on similarities, Clusters are formed by using K-means clustering, then the keyword is searched in clustered instead of several documents. Then the searching is performed in efficient way.