16.Affective recommender systems

Affective recommender systems

Tuning algoritmic parameters

16. Affective recommender systems

The majority of recommender systems are based on ratings from users and this gives a static user-experience as a system. Could recommender systems include other parameters that impact the predictions such as user reaction, behavior, feeling, or emotion. Doing so could create a a more dynamic, customized and tailored experience.

What can designers do with this pattern?


Still to be found in the wild. Very nice post on this from Fanamby Randri: Emotion-based Music Recommendation System Using A Deep Reinforcement Learning Approach

Input or output?


How do users control the algorithm?

It’s purpose is to recommend the most suitable top k songs to the user from a given playlist, based on previous recordings of his feelings or other biometrics.

How do users understand the control?

Input a playlist of your choice and you information (age, mood, and gender) and upload biometrics such as Heart Rate Variability metrics.

Related Patterns

Related algoritmic affordance pattern in this library are listed below.