A fuzzy ontology framework in information retrieval using semantic query expansion

• Main Authors: Shivani Jain, K.R. Seeja, Rajni Jindal
• Format: Article
• Language: English
• Published: Elsevier 2021-04-01
• Series: International Journal of Information Management Data Insights
• Subjects: Semantic Web; Information retrieval; Fuzzy-ontology; Query-expansion; Search engines; Information systems development
• Online Access: http://www.sciencedirect.com/science/article/pii/S2667096821000021

World Wide Web (WWW) constitutes fuzzy information and requires soft computing techniques to deal context of the query. It works on the principle of keyword matching yielding low precision and recall. Semantic web, an extension WWW improves the information retrieval process. Query expansion is utmost importance in information retrieval to retrieve relevant results. To overcome the weaknesses of current web system and to utilize the strengths query expansion a novel framework based on fuzzy ontology is proposed for information retrieval. In the proposed framework, domain specific knowledge is utilized for ontology construction. In framework pre-defined domain ontologies and Global ontology, ConceptNet is used to construct a fuzzy ontology. Based on constructed fuzzy ontology most semantically related words for a query are identified and query is expanded. A fuzzy membership function is defined for different semantic relationships present among the Global ontology ConceptNet.Based on the proposed framework queries are expanded (Semantic query expansion) and evaluated on four popular search engines namely Google, Yahoo, Bing and Exalead. The performance metrics used are Precision, Mean Average Precision (MAP), Mean Reciprocal Rank (MRR), R-precision and Number of documents retrieved. The Web search engines are precision oriented. Based on the proposed framework all the metrics are improved approx. by 10%. Precision before the query expansion lies between 0.75-0.81 whereas after the query expansion lies between 0.85-0.89 on various search engines. The number of documents retrieved is almost improved 1/1000 after the query expansion.