Open theo-o opened 4 years ago
What if we considered the sponsor and/or the sponsor's department as well as the activity itself when making recommendations? For instance, if you frequently attend a Dr. White club, his study hall should be recommended and clubs run by other CS teachers should get higher scores. This is based in the idea that sponsors in the same department would sponsor similar clubs.
Topic of discussion
Currently, we use a fairly simple means of recommending users for activities (https://github.com/tjcsl/ion/blob/master/intranet/apps/eighth/management/commands/generate_similarities.py). This approach is not completely accurate.
Proposed options/solutions
I am not fully versed in all the possible solutions, so a lot of possible approaches exist and I am welcome to hear them.
Personal opinion (if applicable)
The main criteria for a good solution are: