Enhancing the degree of personalization through Vector Space Model and Profile Ontology

A Relevance-Focused Search Application for Personalised Ranking Model

“…The personalisation of search results to a large degree lies in merging the models that provide them. A description of the linear combination adopted in the current research can be found in [ 2]. Here, as outlined in the following section, the aim is to test the system' models on a deeper level and to investigate their real world problems as closely as possible.…”

“…Although there are many overlaps between the current research and the latter approach aimed at providing semantic similarities through ontologies, in terms of classification technique employed to create users' profiles to describe the contents of Web documents clicked, this project applies both term weight (i.e. term frequency factor) and dwell weight 3 directly as a dimensional feature to enrich the users' models [ 1,2]. For instance, not only was it shown in these surveys that the performance of the PRM improved, but it was also demonstrated that it could be used as a complementary feature for the system to rely on when the keyword feature proves unsuccessful in identifying the relevance of documents.…”

“…Here it is important to recall that words which have been integrated are not directly related to the keyword queries, thus, it is not feasible to apply the cosine similarity measure directly. The application of fuzzy ontology values as shown in equation 3.4 [ 2] addresses this problem. Thus, based upon this similarity measure (i.e.…”