Why we use recommender systems?


A recommender system seeks to estimate and predict user content preference regarding games, stories or videos. The system draws from data usage history, aiming at making suggestions based on the user’s (current) interests.  

A recommender system usually employs data sources to learn more about such preferences, making good use of explicit feedback resulting from diverse evaluation metrics such as “Add To Favorites” for example, or implicit feedback deriving from the number and length of content-based interactions.   

Within the implicit feedback approach, a basic algorithm despite being more elaborate than the one which generates completely random recommendations, consists of reflecting most popular content, summing up all user activity and recommending most common content in relation to the number and length of visits.

A third algorithm is about quantifying user preference similarity based on the visited content which the users jointly have, identifying groups by “preference profile”. Once these groups have been identified, the system will recommend to the user content which has not yet been visited, but seems appealing to the other members of the group. This algorithm is known as collaborative filtering.

Lastly, it is possible to employ a hybrid algorithm, that is to say, an algorithm which combines content predictions based on popularity and novelty metrics, and collaborative filtering. This is the method we implement at Smile and Learn, generating recommendation content lists, with regard to each one of the algorithms, so as to rearrange and fuse data, building upon new content prediction results.

It so happens, that the more feedback data-driven decision models can collect, the more robust the prediction will be; that’s why it’s crucial for the content to be highly appealing. As we access new data, we make room for a promising chapter in which we’re able to detect new patterns, develop more explanatory hypotheses and progressively design more informative and consequently relevant, usage and learning metrics.

Jesús Rodriguez

Data Science Team

no replies

Leave your comment