Nowadays, the purchase of every product involves a lot of critical thinking. Every buyer goes through a lot of user reviews and rating before finalizing his purchase. They do this to ensure that the product they purchase is of good quality at minimum price possible. It is evident now that online reviews are not that reliable because of fake reviews and review bots. Now you can even pay certain social media groups to give your product a fake good rating. Hence going just with the reviews of some stranger whom you do not know is not intelligent. So we propose a recommendation model based on the Trust Relations (TR) and User Credibility (UC) because it is human nature that a person feels more comfortable when he gets a review from a person he knows on a first name basis. Also, the credibility of the reviewer is an important factor while providing importance to the reviews because every person is different from other and can have different expertise. Our model takes into account the effect of credibility which is not used by any other recommendations models which increases the precision of the results of our model. We also propose the algorithm to calculate the credibility of any node in the network. The results are validated using a dataset and applying our proposed model and traditional average rating model which shows that our model performs better and gives precise results.
All Science Journal Classification (ASJC) codes
- Statistics and Probability
- Artificial Intelligence