Something else is lurking in the recommendations

Recommender systems are almost in control of our daily choices. The recommendations of films, music and services surround us. But how to create them correctly? It would seem that you took the recommendation algorithm as the basis and wrapped your data, but no! It turns out that not only quality metrics evaluate models, but also many other factors. Let's see what else affects the choice and how this choice can be controlled, regardless of the model metric.

Artem Seleznev

Artem Seleznev MegaFon