May I ask why the recommender task (KGCN) dataset size in the download link provided is different from that in the paper? (Maybe the page showed 495MB, actually only 475 were obtained)
For example:the movie folder get
number of users: 90359
number of items: 16954
number of entities (containing items): 102569
number of relations: 32;
and the music dataset is identical with the original KGCN paper.
So how can I get the new dataset you construct in your paper for recommender task? Thanks
Hi. The latest dataset is under baseline folder or dataset page of our website. The dataset in KGCN folder is the original experiment for KGCN paper. For HGB benchmark performance of KGCN, you can refer to here.
May I ask why the recommender task (KGCN) dataset size in the download link provided is different from that in the paper? (Maybe the page showed 495MB, actually only 475 were obtained) For example:the movie folder get number of users: 90359 number of items: 16954 number of entities (containing items): 102569 number of relations: 32; and the music dataset is identical with the original KGCN paper. So how can I get the new dataset you construct in your paper for recommender task? Thanks