Improving Implicit Recommender Systems with View Data
Authors: Jingtao Ding, Guanghui Yu, Xiangnan He, Yuhan Quan, Yong Li, Tat-Seng Chua, Depeng Jin, Jiajie Yu
IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on two real-world datasets demonstrate that our method outperforms several state-of-the-art MF methods by 10% 28.4%. |
| Researcher Affiliation | Collaboration | Jingtao Ding1, Guanghui Yu1, Xiangnan He2, Yuhan Quan1, Yong Li1, Tat-Seng Chua2, Depeng Jin1 and Jiajie Yu3 1 Beijing National Research Center for Information Science and Technology (BNRist), Department of Electronic Engineering, Tsinghua University 2 School of Computing, National University of Singapore 3 Beibei Inc |
| Pseudocode | Yes | Algorithm 1: Fast VALS Learning algorithm. |
| Open Source Code | Yes | Our implementation is available at: https: //github.com/dingjingtao/View enhanced ALS. |
| Open Datasets | Yes | Tmall2: Tmall is the largest E-commerce platform in China. To make our results reproducible, we use a public benchmark released in IJCAI-2015 challenge3. The dataset is downloaded from https://tianchi.aliyun.com/ datalab/data Set.htm?id=5 |
| Dataset Splits | No | The paper uses a 'leave-one-out protocol' for testing but does not explicitly mention a separate validation split. Training is done on 'the remaining data'. |
| Hardware Specification | Yes | Our experiment on the same machine (Intel Xeon 2.10 GHz CPU) shows that the actual training time per iteration for VALS on Beibei dataset is about 75s. |
| Software Dependencies | No | No specific software versions (e.g., library names with version numbers) were provided. |
| Experiment Setup | Yes | We tuned the weight of missing data si. We tuned the learning rate εBPR. For regularization, we set λ as 0.001 for all methods for a fair comparison. Since the findings are consistent across the number of latent factors K, we report the results of K = 32 only. Our VALS has two parameters: {γ1, γ2}, which are the margin values... and ci that determines the weight of view data. To find the best setting for (γ, c0), we conduct a grid search over these two parameters... we fix γ and c0 according to the best performance evaluated by HR, i.e., γ =0.3, c0 =1.6 for Beibei and γ =3.5, c0 =0.5 for Tmall. |