@techreport{oai:ipsj.ixsq.nii.ac.jp:00211767, author = {田中, 直哉 and 藤井, 章博 and 清水, 宏泰 and Naoya, Tanaka and Akihiro, Fujii and Hiroyasu, Shimizu}, issue = {2}, month = {Jun}, note = {現在,心疾患の特定や突然死リスクの高低度合いについて,心電図データから目視で分類が行われている.本研究では,心電図データから,特徴波形を抽出,それをデータセットとして機械学習によるリスク評価を自動的に行う手法を提案する.また,リスク評価について,複数の機械学習アルゴリズムを比較することにより,提案手法において最適と思われるアルゴリズムの検討も行った., Hear diseases has risk of sudden death.The diagnosis of the diseases is usually performed by visual identification from the electrocardiogram data. In this study, we extract waveforms from electrocardiogram data, then the characteristic of the form is classified by machine learning scheme. Risk evaluation is done automatically based on this classifications. We have proposed several machine learning algorithms in terms of risk assessments and compared them to find out optimal methodology for the data set.}, title = {心電図データにおける特徴抽出}, year = {2021} }