A Domain-Specific Search Engine: A Case of University of Abuja

Abstract

Users need more effective ways of organizing and searching for information. It has become increasingly difficult for users to find
information on the World Wide Web that satisfies their individual needs since information resources on the World Wide Web continue to grow. The reason is that the World Wide Web provides a bunch of information that is generic in perspective. Although, the generic search engine makes attempts to streamline searching to search-key by traversing and combing the whole bunch of stored information. Under these circumstances, the system turns out to be inefficient and many times the results do not tally to what users desire. This paper presents a system to find interesting textual content among tons of documents taking the department of computer science of the University of Abuja, Nigeria as a case study. The paper proposed a vector space ranking algorithm which is a content-based ranking method that allows the user to utilize the full co-occurrence matrix of all words in the corpus to bring out relevant material using a simple and structured query interface. The index structures of the system have been specifically designed to support the ranking scheme. One important aspect of the system is its flexibility and robustness that makes it adaptable to any domain rather than being tailored to a particular one.