VExSearch: Improving visualizations results for web-based information exploration and refinement

As the web grows, finding information from large data repositories has become increasingly difficult not only due to inadequate number of results that are relevant, but also due to poor sorting of relevant results from those irrelevant. The present search engines (SEs) use the query and response lookup process that does not provide precise results. Thus, researchers have gone beyond the paradigm to explore a new class of method to seek information, which is called exploratory research that is open-ended and its faceted search can improve the overall search process. Besides, many studies have begun tapping into enhancement of web search results relevancy. The web reflects vast heterogeneity, varying structure, and massive in volumes. Therefore, it is rather difficult to seek accurate outcomes as desired. As such, visualisation and interactive graphics have been proposed as methods to manage massive amounts of results and to project essential features for the web pages. Additionally, search engine controls reconstruction and reformulation of queries. As such, a search engine is presented in this study by developing it on the cloud computing platform environment The search engine is based on the idea of improving visual exploratory search (VExSearch) while exploring information in the web. This particular notion reflects the process of seeking and combing through the vast information by using the coordinated visualisation method, apart from minimising the effort spent in seeking information per query. The VExSearch was evaluated for its capability and performance and later compared with IMDb SE and CloudMining SE. The comparative results showed that the VExSearch was 66% more accurate than the other SEs. VExSearch also seemed to provide the most relevant results among all the three SEs, aside from attaining an average improvement of 20% in terms of recall.