Title: AN EFFECTIVE ROCOMMENDER MODEL FOR E-COMMERCE PLATFORMS
Abstract: Because of the sparsity problems in databases, fake accounts can easily affect results of recommender algorithms especially when a product does not have enough votes by consumers. Generally, fake accounts are created by the owner of the product in order to raise their product score or by the ill-wishers who wants to denigrate a product or a company. This situation represents a great sense for e-commerce platforms especially when considering that majority of companies have less than 1% density of database. In order to overcome negative effects of the fake accounts in e-commerce platforms, this study proposes a recommender model, which will find the consumers who are trustful and have a great effect on other’s opinion by analyzing the relationship between consumers. With the proposed model, the Recommender Systems (RS) are expected to provide recommendations to customers based on trustful users’ opinions to improve the quality of RS in e-commerce platforms.