A framework for measuring similarity between Terms in Short Text Categorization
Vogirala Nandini, Janani Chitra R., P. Uma Maheswari · 2016
Due to the increase in the information availability on the World Wide Web, it becomes too tough for the search engine to provide the precise results for the user. Some information on the web pages is ambiguous in nature. Semantic similarity is worn to measure the similarity score for the text and it improves the efficiency of the search by obtaining the user query and process them consistent with the searcher's intent and it produces the contextual meaning of terms which generates similar results for the query. The semantic search system considers the position of words, user intention, synonyms and relationship between words to produce the correct results for the user. Generating similarity between two ideas is essential for various applications and those applications are used for producing the user satisfactory results. However, the existing approach is additionally appropriate for semantic similarity between words instead of Multi-Word Expressions (MWE) and they do not scale very well. This paper proposes a clustering and classification algorithm for semantic similarity using sample web pages. Further improvement is to analyze the short text for classification and labeling the short text according to the keyword and producing the result for the end user. This type of classification is suited for opinion mining with the tweets from twitter, topic content discovery etc.