NAIS: Neural Attentive Item Similarity Model for Recommendation

A Non-official Implementation of ”NAIS: Neural Attentive Item Similarity Model for Recommendation”.

If you use the codes for your paper as baseline implementation, please cite the link:

https://github.com/hegongshan/neural_attentive_item_similarity

Official Implemenation (Python 2.7 + TensorFlow 1.x): https://github.com/AaronHeee/Neural-Attentive-Item-Similarity-Model

Requirements

• Python 3

• TensorFlow 2.0+

• NumPy (latest version)

• SciPy (latest version)

Example to run the codes

• FISM

python FISM.py --path data --data_set_name ml-1m --epochs 100 --num_neg 4 --embedding_size 16 --lr 0.01 --alpha 0.0 --regs (1e-7, 1e-7, 1e-7)

• NAIS

python NAIS.py --pretrain 1 --path data --data_set_name ml-1m --epochs 100 --num_neg 4 --embedding_size 16 --lr 0.01

Experimental Results

• FISM

epochs = 100

	HR@10	NDCG@10
ml-1m	0.6526	0.3857

• NAIS

coming soon...

	HR@10	NDCG@10
ml-1m

Last Updated: November 16, 2019