Challenges And Solutions To Providing Responsible Recommendations In Social Media

Recommender systems (RS) are a type of information filtering software tools and techniques with the purpose of predicting the rating or affinity of a user to an item (such as music, books, movies) they had not yet considered, using the characteristics of an item (content based approaches) or the user’s social environment (collaborative filtering based approaches). In recent years, RS have become very common in e-commerce, social media and the Internet in general, being an important part of many users’ personalized browsing experiences. Accordingly, there is a need for generating socially responsible recommendations. The immediate objective of this work is to present ways of generating such recommendations in the context of social media and, to a lesser extent, e-commerce. As a research methodology, a critical survey was made, taking into account several major directions, such as: how the relevance of recommendations can be increased, how privacy-related issues that affect user data based recommender systems can be solved or how malicious recommendations can be eradicated. This work is part of a growing body of work on RS research, which aims at improving the reliability factor of e-recommendations and, implicitly, adding value to various RS stakeholders (users and providers).