메뉴 건너뛰기
소속 기관 / 학교 인증
인증하면 논문, 학술자료 등을  무료로 열람할 수 있어요.
한국대학교, 누리자동차, 시립도서관 등 나의 기관을 확인해보세요
(국내 대학 90% 이상 구독 중)
고객센터 ENG
주제분류

논문 기본 정보

저자정보
(울산대학교) (울산대학교)
저널정보
대한전자공학회 대한전자공학회 학술대회 2024년도 대한전자공학회 추계학술대회 논문집
오류 신고하기

피인용 0

검색

    초록·키워드

    This study proposes a hybrid model, the Sparse Residual LSTM Autoencoder (SRL-AE), combining a Sparse Autoencoder and Residual LSTM to improve anomaly detection in electrocardiogram (ECG) data. Traditional LSTM Autoencoder models distinguish between normal and abnormal data based on recon struction error distributions, but the distance between these distributions is often minimal, leading to reduced confidence in model inference. To address this issue, a model that can robustly separate these distributions is required, particularly in the medical field, where the reliability of deep learning model inferences is critical. The proposed SRL-AE introduces a sparsity constraint in the encoder and residual connections in the decoder to increase the distance between the recon struction error distributions. Experimental results using the ECG5000 dataset demonstrate that SRL-AE reduces the reconstruction error of normal data by 0.908 and increases the distance between the error distributions of normal and abnormal data by 2.224, thereby improving the robustness of the model’s inference.

    최근 본 자료 전체보기