Summary of The Effects Of Data Imbalance Under a Federated Learning Approach For Credit Risk Forecasting, by Shuyao Zhang et al.
The Effects of Data Imbalance Under a Federated Learning Approach for Credit Risk Forecasting
by Shuyao Zhang, Jordan Tay, Pedro Baiz
First submitted to arxiv on: 14 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this study, researchers explored the application of Federated Learning (FL) in credit risk forecasting, a crucial task for commercial banks. FL allows for global model training without accessing sensitive client information directly, mitigating security and privacy risks. The team investigated the effects of data imbalance on model performance using three datasets and various scenarios involving different numbers of clients and data distribution configurations. They compared the performance of neural network architectures (MLP and LSTM) and a tree ensemble architecture (XGBoost). Results show that federated models consistently outperform local models for non-dominant clients with smaller datasets, yielding an average improvement of 17.92% in model performance. However, for dominant clients, federated models may not exhibit superior performance, suggesting the need for incentives to encourage participation. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Credit risk forecasting is important for banks and financial institutions. Traditional machine learning methods can be risky because they require sharing sensitive client information. Federated Learning (FL) is a new technique that allows global model training without accessing private data directly. Researchers tested FL in credit risk assessment and found out how it works with different types of data and models. They used three datasets and different scenarios to see how well the models performed. The results show that FL can help banks make better decisions about lending. |
Keywords
* Artificial intelligence * Federated learning * Lstm * Machine learning * Neural network * Xgboost
