Author Email: email@example.com
Purpose of this paper is to resolve the problem of traditional search engines which hardly provide the essential content relevant to the user query. Current techniques for processing of query are mostly based on keywords. Thus, they have limited capabilities to understand the concepts and meaning involved in the user query. The solution for this is a semantic information search. A semantic search overcomes the drawbacks of mismatch associated with traditional keyword-based search. We create a semantic information retrieval system by combining Natural Language Processing and Graph Analysis. The idea of higher-level conceptual understanding of queries is developed to overcome the limitation of the traditional model. We present the method for semantic information retrieval by creating a semantic graph of the query. We then add synonyms and hyponyms of the query terms present in the semantic graph to retrieve the accurate documents. The level of accuracy will be enhanced since the query is analyzed semantically.
Current query processing techniques are depends on keywords. Thus, it has very few ways to grasp the concepts present in the user query. The higher-level conceptual understanding of queries is the main idea of this paper. As traditional information retrieval system is mainly based on keywords thus it is simple and easy. However, traditional system is based on the simple syntax matching with lack of understanding, leading to unsatisfactory search quality and results. Even though the query processing system has various enhancements, but have various limitations in terms of searching as it searches based on literal things. To overcome the limitations of the traditional model i.e. keywordbased model, we introduce a semantic search [1, 2] for information retrieval. A semantic search overcomes the limitation of conceptualization associated with keyword-based search. The traditional information retrieval system based only on the occurrence of words in a document and ignores all other words associated with the keyword. For example, synonyms, hyponyms, etc. For semantic search, we used the graph-based
method. The graph-based method used concept hierarchy and also supports the logical inference of the query. Semantic information retrieval has become an important part of IR field. The natural language processing (NLP)  method gives us the ability to represent keywords in terms of different part of speech. Using this we can find the relationship between various terms in the user query. By creating the semantic graph we can find the level of the graph and give weights to them for categorizing important terms. Addition of synonym and hyponyms finds more accurate documents. In the keyword based system, synonym and hyponyms are not taken into consideration thus some related document is not retrieved. The proposed method gives us the ability to semantically retrieve the relevant document.
In this paper, we have presented a semantic information retrieval model which, extends the traditional information retrieval model, addresses the challenge of conceptualization, and integrates the advantages of both semantic-based and keyword-based information retrieval. By understanding contextual meaning and searcher intent, semantic search improves search accuracy. For the various domains of information search like mobile, vehicle, books, medical, government information systems, etc., the proposed system will be useful. The result of evaluation has shown that the semantic information retrieval model performs much better than traditional keyword-based information retrieval model. The accuracy and retrieval of the relevant document are improved in the semantic model. Future research will aim to present the higher level of semantic search, like one can run this system for World Wide Web to access relevant documents by adding various domains into the system, with the last aim to provide higher level of query understanding when answering user’s needs and retrieve accurate documents.
332 total views, 1 views today