Ontology-Based Full-Text Searching Using Named Entity Recognition

Abstract

Travelling to different places depends on lots of factors such as hotels, restaurants, nearby hospitals, places to visit in cities, etc. All this information is available on different websites in an unstructured manner thus people do not get information as per their queries in organized format. People search for these factors on search engines which use keyword matching mechanism. Therefore, this paper presents full-text queries searching mechanism which gives precise results in a structured format. Here, our system scraps data from websites to collect information about cities, hotels and hospitals. Concepts of linked data using ontology are implied which has the capability to relate multiple sources of data available on different websites and infer new knowledge from it. Natural Language processing methods such as co-reference resolution is used, which forms a relationship between sentences scrapped from web, which helps to perform better search query without losing meaning of sentences during the processing. In our work, we have also used the Named entity recognition mechanism which applies tags on words with the real-world concepts that they represent. These tags are further utilized by Python library named RDFLib to match the tags which form a relationship between classes within ontology. This relationship between classes and tags are further used to insert and extract data from ontology.