基于BP神经网络的睡眠呼吸综合症智能检测系统

提出了应用BP网络对睡眠呼吸暂停病症进行检测的技术方法。考虑到临床检测采集具有数据量大，伴有噪音干扰等特点，采用BP网络，其并行性、客错性较好，具有较强的自适应性及通过实例学习的能力，有助于克服单个诊断医师在知识获取方面存在的“瓶颈”问题。开发出来的睡眠呼吸综合症智能检测系统可以选择采集好的病人样本数据进行自学习，建立病人的特征模型，自动进行数据分析和异常情况的识别与分类，对病人的患病率进行预测。通过实验测试，该系统预测的准确率在88．5％左右，高于单纯的人工诊断，并显著提高了诊断效率。

INTELLIGENT DETECTION SYSTEM OF SLEEP APNEA SYNDROME BASED ON BP NETWORK

The method of detecting the Sleep Apnea Syndrome by back propagation BP network is presented. Considering there is a huge amounts of sampling data with the noise in diagnoses, the BP network is adopted because of its adaptive learning, high parallel and robust to error. It can overcome the bottle-neck problem of knowledge acquirement for single doctor in diagnoses. The system can learn adaptively from Sampling characteristic data of patient, create the illness model, analyze the data, recognize and classify the pattern of syndrome automatically to predict the probability of illness. From the result of clinical experiment, it demonstrates that the accuracy of prediction is 88.5% , higher than the accuracy of manual diagnoses, and improving the efficiency remarkably.