Hi everyone,
If we want to evaluate the results generated by a context-aware recommender systems, which among the existing metrics would be suitable for this situation. Also, i would like to know about relationship between context aware metrics and other existing metrics, if any.
Regards
Zafar Ali
Dear Zafar,
There are many different metrics used in recommender systems to evaluate your results. The most popular are 1) Precision-Recall, 2) NDCG, 3) RMSE, 4) MAE, 5) AUC (Roc). Some of them measure the accuracy, other measure the relevance, and other measure the error significance.
Dear Pavlos Kefalas,
Thanks for your time and good answer. As we know that these metrics are used to evaluate the results of traditional recommender systems (i.e. CB and CF systems). If we want to evaluate the performance of context-aware recommender systems then which among the available evaluation approaches can be used. Also, if we want to incorporate context-awareness in our system then what would be its impact on other dimensions such as accuracy, novelty, serendipity, diversity, privacy etc.
Thanks
You can see this paper:
Comparing context-aware recommender systems in terms of accuracy and diversity by Umberto Panniello, Alexander Tuzhilin, Michele Gorgoglione.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis