Link:
Publisher:
| International Conference on Knowledge Science, Engineering and Management |
Abstract:
| People are checking different sites before doing their business to either purchase any item online or select any service or product. Many commercial sites rely on the reviews to evaluate their product and services. Other sites are especially designed for the users and reviews to e valuate any product or service. However, select the best review is still a big challenge for the user to select. Many works have been proposed to select the best reviews but with contain redundant information. The best personalized review that is really related to the main topic that the user is searching on is very important. For this reason, in this work, a new personalized reviews’ selection is proposed. We based on the idea of that different point of view for the user causes different evaluation and revering. For this reason, searching on the best reviews in a specific subject gives more accurate and significant selection results. In this paper, design a new approach for the best personalized reviews’ selection that is based on two stages. The first one in the predict the subject aspect modeling (distribution) based on using the A latent Dirichlet allocation (LDA) model. Second, we design a new weighted personalized reviews selection based subject aspect scoring function to select the top personalized reviews. The experimental results show our method selects reviews that are more focusing on the product or service subject aspect. The reviews that are more emphasizing on a different subject are selected. |