these are the codes of my project in lab.
Basing on it, I ran experiments and wrote a paper.
I use python to handle part of the data processing job, including:
- split the data into training set and testing set
- build the social network basing on the social data and run BFS on it to search friends for each user
- plot results
part of the code will rely on the extra lib "SNAP", so you should copy the "snap.py"
and "_snap.pyd" intofiles such as baiduPy
I use matlab to implement several traditional and state of the art recommendation methods, including:
1.User based CF
2.Item based CF
3.SVD
4.SocialMF
5.CF-ULF / PureTrust / Trust-CF-ULF (three novel user based CF methods)
6.Self organizing maps (SOM) neural network
7.my own algorithms