Designing Personalized Recommendation in E-Commerce Site Based on Content-Based and Collaborative Filtering


After the great development in the use of the internet and widespread use of e-commerce sites, the problem of the large number of products offered appeared, thus causing confusion for customers in choosing the suitable products for their needs. To overcome this problem, e_commerce sites using the recommendation systems to nominate products to customers personally according to their needs. The proposed system combines two types of recommendation systems they are content-based filtering which uses cosine similarity function to find the similarities between the products described by texts and collaborative filtering, which uses correlation similarity function to nominate products similar in the evaluation to the products that have been previously purchased and evaluation by customers.Several problems are solved by the proposed system in a simple way and at the same time with high efficiency and accuracy, such as the clod start, scalability, synonymy and spared the data problem.