Application of LightGBM hybrid model based on TPE algorithm optimization in sleep apnea detection

• DOI: 10.3389/fnins.2024.1324933
• Publication Date: 2024-02-19T04:27:47Z
• Authors: Xin Xiong; Aikun Wang; Jianfeng He; Chunwu Wang; Ruixiang Liu; Zhiran Sun; Jiancong Zhang; Jing Zhang
• Keywords: Sleep; Gold standard (test)

Introduction Sleep apnoea syndrome (SAS) is a serious sleep disorder and early detection of not only reduces treatment costs but also saves lives. Conventional polysomnography (PSG) widely regarded as the gold standard diagnostic tool for apnoea. However, this method expensive, time-consuming inherently disruptive to sleep. Recent studies have pointed out that ECG analysis simple effective apnea, which can effectively provide physicians with an aid diagnosis reduce patients’ suffering. Methods To end, in paper proposes LightGBM hybrid model based on signals efficient apnea. Firstly, improved Isolated Forest algorithm introduced remove abnormal data solve sample imbalance problem. Secondly, parameters are optimised by TPE (Tree-structured Parzen Estimator) determine best parameter configuration model. Finally, fusion TPE_OptGBM used detect In experimental phase, we validated database provided Phillips-University Marburg, Germany. Results The results show proposed achieves accuracy 95.08%, precision 94.80%, recall 97.51%, F1 value 96.14%. Discussion All these evaluation indicators better than current mainstream models, expected assist doctor’s process medical experience patients.