A Semantic Search Engine using Semantic Similarity Measure Between Words

Measuring semantic similarity between words is very useful in information retrieval. Semantic similarity measure is so useful in many applications, and in the proposed work it is used to create a model Semantic Search Engine. The Semantic Search Engine uses in one hand a Technical Database for computer technology and a Semantic Similarity database to retrieve the resultant Web page for the query word. When a query word is given in the user interface the search engine first searches for the word in the technical database if the word is present the respective Webpage is displayed. If the word is not present in the technical database then the query word is searched in the semantic similarity database. If there are any similar words for the query word those words are displayed as recommendations to the user. The user has to select one of the similar words from the recommendation and accordingly the result page is retrieved. The semantic similarity measure between the words is evaluated using both Pearson correlation coefficient and Spearman correlation coefficient. The time taken to retrieve the relevant Webpage in semantic search engine is compared with normal search engine. The Precision and Recall is calculated for semantic search engine and the results are compared with normal search engine.