Diversifying Personalized Recommendation with User-session Context

Authors: Liang Hu, Longbing Cao, Shoujin Wang, Guandong Xu, Jian Cao, Zhiping Gu

IJCAI 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experiments on a real-world e-commerce dataset show the superiority of our model over the state-of-the-art methods.
Researcher Affiliation Academia University of Technology, Sydney Shanghai Jiao Tong University *Shanghai Technical Institute of Electronics & Information
Pseudocode Yes Algorithm 1 A GPU-based SGD Optimizer for SWIWO
Open Source Code Yes The MATLAB implementation1 of Algorithm 1 is provided online for more details. 1https://github.com/rainmilk/ijcai17swiwo
Open Datasets Yes We used the IJCAI-15 competition dataset2 for our experiments. This real-world dataset was collected from Tmall.com... 2https://tianchi.shuju.aliyun.com/datalab/dataSet.htm?id=1
Dataset Splits No From the six-month shopping logs, we randomly held out 20% of the sessions from the last 30 days for testing, and the remaining data are used for training. It does not explicitly mention a separate 'validation' split, only training and testing.
Hardware Specification No Not found. The paper mentions a 'GPU-based' optimizer but provides no specific details about the hardware (e.g., specific GPU models, CPUs, memory).
Software Dependencies No The MATLAB implementation1 of Algorithm 1 is provided online for more details. Our experimental results were obtained using Adam. No version numbers are specified for MATLAB or the Adam optimizer.
Experiment Setup Yes We set 50 units for the context embeddings and 10 units for the user embeddings when training SWIWO-I and SWIWO models.