Open rituk opened 2 years ago
Hello? any suggestions?
Hi, thanks for your interests. We directly use the existing datasets from datasets. However, these datasets are anonymized, so it may be hard to identify their real attributes, such as names and categories. But you can generate the recommendation predictions based on the test datasplit to evaluate the prediction accuracy with groud-truth interaction records.
I see, okay. thank you for getting back to me.
Are there any example recommendations generated for movie/music/book dataset?