Φίλτρα συνεργασίας σε περιορισμένα κοινωνικά δίκτυα
Collaborating filtering in a limited social network
The creation of a wide network that is today known as the internet gave commercial corporations the solid base to look for customers. This led the corporations to focus in the development of websites and web applications that have as main purpose to present the merchandisable products and the services to the users of that network. Soon it became a fact that the simple presentation of those products is not good enough to convince the users to prefer their products from the growing completion’s similar ones. The recommendation systems took their part to help find the solution to such problem and attract a larger amount of users. Recommendations systems main goal is to provide users a more targeted amount of suggestions. This way users are provided with products that are more suitable to them, something that increases the probability of them purchasing those products. The modern trend shows an increasing interest of users in using social media networks. As a result an increased number of those networks were created in which users are not an individual entity but members of a wide group. A user that belongs to such a network is more prone to revealing his consumer preferences in products and services. This user is always seeking his connected fellow users and takes their opinion rather seriously. The main purpose of this diploma thesis is to create a recommendations algorithm which is based on user interaction with products. In addition with this algorithm we will create a limited social network in which the users that belong to it will be a part of it and interact with the algorithm. The result of the combination of the recommendations algorithm with that limited social network will create a unified recommendations system that will include the advantages of both those technologies in order to maximize the quality of the recommended items.