A Tunable and Explainable Attributes (TEA) for Recommendation System

Recommender system in information retrieval systems recommends relevant items to consumers by inspecting the consumer's preferences and objective behaviours. We received recommendations daily regarding what to take in as food, buy, and choices about the wear trends. However, we can understand our recommendations, and even less, we can tune our preferences efficiently. One main concern is those undesired or irrelevant recommendations, not only wasting our time, but they also cost media companies millions of dollars every year. Considering this issue, we proposed a tunable and explainable recommendation system where the attributes that determined the recommendations are explainable and tunable for every individual consumer. Irrelevant recommendations waste the consumer's time and the company's money, but if we put the power of the recommendations in the consumer's hands, we have a win-win situation. The suggested method addresses recommendation algorithms issues, which creates negative feedback loops and allows consumers to be aware of the profile they are building and tune it to see the content that interests them most.