Caser

A Matlab implementation of Convolutional Sequence Embedding Recommendation Model (Caser) from paper:

Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding, Jiaxi Tang and Ke Wang , WSDM '18

Note: I strongly suggest to use the PyTorch version here, as it has better readability and reproducibility.

Requirements

Matlab R2015 +
MatConvNet v1.0

Usage

Installing MatConvNet (guide).
Change the code to make the path point to your MatConvNet path.
Open Matlab and run main_caser.m

Configurations

Data

Datasets are organized in 2 seperate files: train.txt and test.txt
Same to other data format for recommendation, each file contains a collection of triplets:

user, item, rating

The only difference is the triplets are organized in time order.
As the problem is Sequential Reommendation, the rating doesn't matter, so I convert them to all 1.

Model Args (in main_caser.m)

L: length of sequence
T: number of targets
rate_once: whether each item will only be rated once by each user
early_stop: whether to perform early stop during training
d: number of latent dimensions
nv: number of vertical filters
nh: number of horizontal filters
ac_conv: activation function for convolution layer (i.e., phi_c in paper)
ac_fc: activation function for fully-connected layer (i.e., phi_a in paper)
drop_rate: drop ratio when performing dropout

Citation

If you use this Caser in your paper, please cite the paper:

@inproceedings{tang2018caser,
  title={Personalized Top-N Sequential Recommendation via Convolutional Sequence Embedding},
  author={Tang, Jiaxi and Wang, Ke},
  booktitle={ACM International Conference on Web Search and Data Mining},
  year={2018}
}

Comments

For easy implementation and flexibility, I didn't implement below things:

Didn't make mini-batch in parallel.
Didn't make the model in MatConvNet wrapper.

License

GNU Lesser General Public License v3.0

wenbella / caser