基于支持向量机的酗酒脑电信号分类研究

Research on Classification of Alcoholism EEG Signals Based on Support Vector Machine

酗酒会对脑认知功能产生严重损伤。为了检测长期饮酒人员是否有酗酒倾向问题,提出一种基于支持向量机的酗酒脑电信号特征分类识别方法。借助三类评估参数,实现了对健康者和酗酒者脑电信号定量分析。研究结果发现,能量评估参数是一种新的最佳评估分类因子,采用θ能量评估参数利用支持向量机能够对两类不同特征的脑电信号进行分类识别,识别精度最高达到90%。该方法能够为长期饮酒人员是否有酗酒倾向提供一种新的辅助检测方法,有助于提醒长期饮酒群体对健康给予关注。

Alcohol abuse can cause serious damage to brain cognitive function.In order to detect whether long term drinkers have a tendency to alcohol abuse,a feature classification and recognition method of alcoholic EEG signals based on support vector machine(SVM) is proposed.Quantitative analysis of EEG signals for healthy and alcoholics people is realized by using three kinds of evaluation parameters.The results show that θ-energy evaluation parameter is a new optimal classification factor.With theta-energy evaluation parameter,by using support vector machine,two kinds of EEG signals with different characteristics can be classified and recognized,and the recognition accuracy can reach 90%.This method can provide a new auxiliary detection method for the tendency of alcoholism of long-term drinkers,and is helpful for reminding long-term drinkers pay attention to health.