VOLUME XX, 2017
1
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 10.1109/ACCESS.2017.Doi Number
Federated Learning for Clinical Event
Classification Using Vital Signs Data
Ruzaliev Rakhmiddin
1
, KangYoon Lee
1*
1
Department of Computer Engineering, Gachon University, Fort Collins, SongNamSi, 13120 South. Korea
*
Corresponding author: KangYoon Lee (e-mail:
keylee@gachon.ac.kr
).
This work was supported in part by the Commercialization’s Promotion Agency for R&D Outcome (COMPA) Grant funded by the
Korean Government
(MSIT) under Grant 2022-Future research service development support-1-SB4-1, and in part by the National Research Foundation of Korea (NRF) Grant
funded by MSIT under Grant NRF-2022R1F1A1069069.
ABSTRACT
Although the healthcare industry has advanced with machine learning techniques, big data is
needed to get accurate and fast diagnostic results. The main use of the above offer is to share privacy issues
with others, making effective forecasting d featured applications that require large amounts of training data.
Nowadays, Federated learning is used to keep the privacy of the data by using the main server with several
clients. In this study, we propose clinical event classification using vital signs data with Federated Learning.
Here, the datasets of X hospital clients with vital signs are using several types of machine learning regression
models that leverage your own data locally such
as Random Forest Regression,
XGBoots Regression,
Logistic Regression etc. Our method got a high positive result in terms of accuracy with 98.9 percent while
the classification of clinical event.