基于粒子群优化的支持向量机用于直肠感知功能重建

针对临床上肛门失禁导致的直肠感知功能丧失,提出了一种基于粒子群优化(PSO)的支持向量机(SVM)重建患者直肠感知功能的方法.分析人体直肠压力生理特征,将典型直肠压力收缩波形中的巨大移行性收缩(HAPC)作为产生便意的主要依据,利用小波包分析对直肠压力信号进行特征提取,通过提取的特征向量对基于SVM的直肠感知预测模型进行训练,使用PSO算法对SVM的参数进行优化,并利用训练后的模型进行便意预测,同时对比分析了参数优化后的SVM和不同核函数的SVM便意预测的准确率.实验结果表明,所提出方法切实有效,能够帮助患者重建直肠感知功能.

Rectal Perception Function Rebuilding Based on SupportVector Machine Optimized by Particle Swarm Optimization

Particle swarm optimization (PSO) optimized support vector machine (SVM) based rectal perception function rebuilding method was proposed for rectal perception loss caused by anal incontinence. By analyzing human rectum characteristics, high amplitude propagated contractions (HAPC) in rectal contractions were used to indicate an urge to defecate. Rectal pressure feature was extracted using wavelet packet analysis, taking normalized of wavelet packet coefficients mean and energy as feature vector. Rectal perception prediction model was trained based on SVM whose parameters are optimized by PSO. Then the trained model was used to predict the urge to defecate. And the prediction accuracy of the optimized and non optimized SVM with different kernel functions was compared. Experimental results show that the proposed method is effective in rebuilding patients’ rectal perception function.