A Survey Paper on Existing Recommender Systems

Recommender systems apply data mining techniques and prediction algorithms to predict users' interest on information, products and services among the tremendous amount of available items. The vast growth of information on the Internet as well as number of visitors to websites add some key challenges to recommender systems. Selling a wide variety of products has become easier since the coming of online stores, online retailers are able to sell more products than a physical store. The disadvantage is that the customer is not able to find products anymore, because they have to browse in many different categories and sub-categories in order to find the products they are looking for. That is why the recommender systems are increasingly used on e-commerce websites. It learns from the knowledge about the customers and products and gives suitable personalized recommendations to every single customer. Although recommender systems help discovering products, they do have some disadvantages that have to be considered. Some examples are lack of personalization, inaccurate recommendations or no recommendations at all when there is not enough information about the customer or about the product.
In this paper I will look at different approaches that can be used for recommendations. The main question in this paper is what approach will give the best recommendations for an e-shop. These approaches are surveyed using different existing recommender sites. My answer to this question is that it depends on the concrete use case. No matter how clever an approach is, there is no single approach that will give the best recommendations for every retailer, because online retailers and shoppers vary way too much.