An Improved Web Explorer using Explicit Semantic Similarity with ontology and TF-IDF Approach

The Improved Web Explorer aims at extraction and selection of the best possible hyperlinks and retrieving more accurate search results for the entered search query. The hyperlinks that are more preferable to the entered search query are evaluated by taking into account weighted values of frequencies of words in search string that are present in anchor texts and plain texts available in title and body tags of various hyperlink pages respectively to retrieve relevant hyperlinks from all available links. Then the concept of ontology is used to gain insights of words in search string by finding their hypernyms, hyponyms and synsets to reach to the source and context of the words in search string. The Explicit Semantic Similarity analysis along with Naïve Bayes method is used to find the semantic similarity between lexically different terms using Wikipedia and Google as explicit semantic analysis tools and calculating the probabilities of occurrence of words in anchor and body texts .Vector Space Model is being used to calculate Term frequency and Inverse document frequency values, and then calculate cosine similarities between the entered Search query and extracted relevant hyperlinks to get the most appropriate relevance wise ranked search results to the entered search string