A Horizontal Federated Random Forest for Heart Disease Detection from Decentralized Local Data
Authors:
- 16 September 2022
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 191-196
Abstract
In the modern world, reliable data is a thriving need in every sector. As the data increases, maintaining data privacy is also becoming a big concern. The healthcare sector is no different than that. Privacy in the healthcare sector is a topmost concern when sharing with other institutes. As data from a single healthcare institute is not always enough to get properly predicted outputs in machine learning approaches. There comes the idea of sharing data among multiple hospitals for having a more specified model with keeping the data details private. So, we have designed a model combining a federated central model and clients for the application of Federated Learning on heart disease patients' data. Here, we have implemented an approach for sharing only the model parameters among the clients and central in horizontal federated learning infused with random forest. At the evaluation of our model, we have come up with improved accuracy of 7.1, 2, and 6 percent respectively for the federated central and both clients.Keywords
This publication has 6 references indexed in Scilit:
- Advances and Open Problems in Federated Learning2021
- Patient clustering improves efficiency of federated machine learning to predict mortality and hospital stay time using distributed electronic medical recordsJournal of Biomedical Informatics, 2019
- Hyperparameters and tuning strategies for random forestWIREs Data Mining and Knowledge Discovery, 2019
- Federated learning of predictive models from federated Electronic Health RecordsInternational Journal of Medical Informatics, 2018
- The SAGE Handbook of Regression Analysis and Causal Inference2014
- Differential Privacy2006