论文部分内容阅读
本文给出了一种基于统计推断识别运动肌无动作电位的新方法(probabilisfic inference based technique,IBC)。该方法的关键思想是:借助一个具有某种先验知识的训练样本集,从中找到统计关系,进而导出分类识别准则。值得指出的是训练样本的要求不高,它可以具有一定程度的非一致性及信息缺失等。文中首先推导了算法,然后将这一算法应用到从模拟肌电信号中抽取运动肌无动作电位的识别中,接着对同样的数据,把这一算法的性能与常用的终极匹配算法(TBC)作了比较,采用32个训练样本作为算例。作者发现IBC
In this paper, we present a new probabilistic inference based technique (IBC) based on statistical inference. The key idea of this method is to find the statistical relationship by using a set of training samples with some prior knowledge, and then derive the classification criteria. It is worth noting that the training sample is not demanding, it can have a certain degree of inconsistency and lack of information. In this paper, the algorithm is first deduced, and then the algorithm is applied to the recognition of the no-action potential of the motor muscle extracted from the simulated myoelectric signals. Then, for the same data, the performance of this algorithm is compared with the commonly used algorithm of ultimate matching (TBC) For comparison, 32 training samples are used as examples. The authors found that IBC