Machine Learning-Based Diagnosis of Epilepsy in Clinical Routine: Lessons Learned from a Retrospective Pilot Study

被引：1

作者：

Rieg, Thilo ^{[1
]}

Frick, Janek ^{[1
]}

Buettner, Ricardo ^{[1
]}

机构：

[1] Aalen Univ, Aalen, Germany

来源：

INFORMATION SYSTEMS AND NEUROSCIENCE, NEUROIS RETREAT 2020 | 2020年 / 43卷

关键词：

Electroencephalography; Random forests; Spectral analysis; Machine learning; SYSTEMS;

D O I：

10.1007/978-3-030-60073-0_32

中图分类号：

B84 [心理学];

学科分类号：

04 ; 0402 ;

摘要：

In this work-in-progress paper, we present preliminary results of a large pilot study for implementing a novel machine learning approach presented at HICSS 2019 [1] and ICIS 2019 [2] in a German hospital to detect epileptic episodes in EEG data. While the algorithm achieved a balanced accuracy of 75.6% on real clinical data we could gain valuable experience regarding the implementation barriers of machine learning algorithms in practice, which is discussed in this paper. These lessons learned have practical implications for future work.

引用

页码：276 / 283

页数：8