Published Paper Details:

In keyword search it is important to identify the actual meaning of a query and the context of query terms at an earlier phase. Conceptualization is the process of identifying this contextual information about query terms and to produce meaningful segments which guides to achieve search diversification. Because of the advancements in Machine Learning (ML), Artificial Intelligence (AI) and the growth in overall Parallel computing applications the search diversification process is not at all a risky job now. The proposed Semantic Search Diversification Model (SSDM) identifies the context of the search query at the very beginning of the search. The concept of Smallest Lowest Common Ancestor (SLCA) nodes is used to lock the desired key and semantic keys over the XML tree branches. The Mutual Information value of the given query terms and their semi features are considered. The Probabilistic weight of the search query is calculated based on the relevancy and the novelty of the query. The proposed Anchor based Semantic Preservation (AESP) algorithm extracts the weighted SLCA nodes for the original and generated queries from the given XML data and determines the anchor nodes to initiate the tree pruning. The node lists are assigned to multiple processors in order to achieve parallelism. The fuzzy set of the popular terms in the given data set is used for creating semantic table for the query terms. The Normalized Discounted Cumulative Gain (nDCG) measures of the SSDM shows the usefulness of the query suggestions, and a point to apply search annotations if needed.