Word Vocabulary of Chosen Datasets about md-cvae

[yaochenzhu]

Hi @trannhuthuat96,

Thank you for your interest in our work, and I'm so sorry for the late response.

It seems that currently, I cannot find the processed word vocabulary. But for the movielen-sub datasets, I have released the processed raw text of the movie plot that I crawled from IMDB (see [Google Drive]), which can be transformed into the item features. As for the citeulike-a dataset, maybe you could try to find the vocabulary from CDL or CVAE, because I use the citeulike-a datasets from their repos. I'm sorry for the inconvenience for your work.

As for the features.npz file, I just calculate the element-wise maximum of the normalized word counts of the observed items for each user as the user features. The details can be found in section 4.1 of my other paper VBAE.

Good luck with your work :)

Best,
Yaochen