-
Notifications
You must be signed in to change notification settings - Fork 225
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
why the fm model didn't consider the data value? #35
Comments
The implementation is proper. A more interesting question could be: "How does this implementation work with continuous features and multi-hot vectors?". In particular, I am interested in the multi-hot question. I read the implementation, which assumes the dataset comes as user-item indexes followed by the ground truth. However, if I want to add movie genres as multi-hot features, how can I do that with this implementation? If we assume each movie has just one genre, it is straightforward since it is enough to add a column after the item index column. If, instead, we have multiple genres for each movie, how is it possible to model that with the current implementation? To reformulate better, an example in the dataset could be encoded as follows, following the syntax of Rendle: Clearly, the item genres are encoded as multi-hot vectors. I want to model this data with factorization machines using the implementation provided in this repository. Thank you very much for your time and effort in realizing such a suitable repository. |
In FeaturesLinear
The model only assign weight value for each field as embedding, but it didn't consider the data value here.
For example, if we has a data as [[1,0]], the model will work embedding("1")+embedding("0"), but in real, we need work embedding("1")*1+embedding("0")*0.
What reason the model delete the effect of data value?
The text was updated successfully, but these errors were encountered: