The present invention relates to an apparatus for collaborative filtering-based product recommendation in a product recommendation service system. In the product recommendation service system, evaluation information and product related information about a product are collected from at least one terminal device, and the collected evaluation information and By using the product-related information, weighting the product evaluation by user through gradient-based collaborative filtering, recommending one or more products, and assigning these weights to users who have a more reliable evaluation can be expected. The accuracy of the evaluation can be increased, and different evaluation criteria can be reflected for each user, thereby increasing the accuracy of recommendation for the product.
展开▼