I added the novelty metric.
Regarding metric definition, I based on the popular paper of Zhou et al. "Solving the apparent diversity-accuracy
dilemma of recommender systems" and the survey paper "How good your recommender system is? A survey on evaluations
in recommendation"
Check the section D2 from Zhou paper in order to evaluate if I understood and implemented right the metric.
PS. Also, I added the novelty in the init .py where I had added the prediction_coverage, there is the conflict because I haven't checked the last PR which you accepted yesterday.
I added the novelty metric. Regarding metric definition, I based on the popular paper of Zhou et al. "Solving the apparent diversity-accuracy dilemma of recommender systems" and the survey paper "How good your recommender system is? A survey on evaluations in recommendation" Check the section D2 from Zhou paper in order to evaluate if I understood and implemented right the metric.
PS. Also, I added the novelty in the init .py where I had added the prediction_coverage, there is the conflict because I haven't checked the last PR which you accepted yesterday.