Recommender systems based on hybrid models

Abstract

[EN]Recommender Systems (RSs) play a very important role in web navigation, ensuring that the users easily find the information they are looking for. Today’s social networks contain a large amount of information and it is necessary that they employ mechanism that will guide users to the information they are interested in. However, to be able to recommend content according to user preferences, it is necessary to analyse their profiles and determine their preferences. The present study presents the work related to different recommender systems focused on two different hybrid models. Both of them are using a Case-Based Reasoning (CBR) system combined with the training of an Artificial Intelligence (AI) algorithm. First, some information is analyzed and trained with an AI algorithm in order to determine relevant patters hidden on the information. Then, the CBR system extends the system using a series of metrics and similar past cases to decide whether the recommendation is likely to be recommended to a user. Finally, the last step on the CBR is to propose recommendations to the final user, whose job is to validate or reject the proposal feeding the cases database.