Financial Risk Analysis for SMEs with Graph-based Supply Chain Mining
Authors: Shuo Yang, Zhiqiang Zhang, Jun Zhou, Yang Wang, Wang Sun, Xingyu Zhong, Yanming Fang, Quan Yu, Yuan Qi
IJCAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on real-world financial datasets prove the effectiveness of our proposal for financial risk analysis for SMEs. |
| Researcher Affiliation | Industry | Shuo Yang , Zhiqiang Zhang , Jun Zhou , Yang Wang , Wang Sun , Xingyu Zhong , Yanming Fang , Quan Yu , Yuan Qi Ant Financial Services Group {kexi.ys, lingyao.zzq, jun.zhoujun, zuoxu.wy, sunwang.sw}@antfin.com, {xingyu.zxy, yanming.fym, jingmin.yq}@mybank.cn, {yuan.qi}@antfin.com |
| Pseudocode | No | The paper does not contain any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | No | The paper states: 'The data used in this research were all processed by data abstraction and data encryption, and the researchers were unable to restore the original data.' and 'Sufficient data protection was carried out during the process of experiments to prevent the data leakage and the data was destroyed after the experiments were finished.' These statements imply that code or data related to the research is not publicly shared for security and privacy reasons. |
| Open Datasets | No | The paper states: 'The SME graph as well as the labeled data for supply chain mining are from Alipay, a mobile cashless payment service. The lending data is from Ant SME Lending, an online credit loan service for SMEs.' and 'The data is only used for academic research and sampled from the original data, therefore it does not represent any real business situation in Ant Financial Services Group.' This indicates the datasets are proprietary and not publicly accessible. |
| Dataset Splits | Yes | Table 2: Statistics of the Datasets. Supply Chain Mining: Train 904K, Validation 301K, Test 602K. Default Prediction: Train 529K, Validation 217K, Test 307K. |
| Hardware Specification | No | The paper states: 'Models are trained on a cluster of 15 Dual-CPU server with AGL framework [Zhang et al., 2020].' This description is too general and does not provide specific CPU models, GPU types, or detailed hardware specifications. |
| Software Dependencies | No | The paper mentions: 'We implement all GNN models in Tensor Flow with the Adam optimizer [Kingma and Ba, 2014]' and 'trained on a cluster of 15 Dual-CPU server with AGL framework [Zhang et al., 2020]'. However, it does not specify version numbers for TensorFlow, Adam optimizer, or the AGL framework, which is required for reproducibility. |
| Experiment Setup | No | The paper states: 'We implement all GNN models in Tensor Flow with the Adam optimizer [Kingma and Ba, 2014], and set the hyperparameters according to the best result in validation set. All the GNN-based models are set to involve 2-hop neighbors.' While it mentions hyperparameters are used and tuned, it does not provide specific values for these hyperparameters (e.g., learning rate, batch size, epochs, specific optimizer settings). |