五、分类符训练为了对机器的工况有更加全面的了解,需要有更多的、具有代表性的样本。这些样本可以取自不同的机器状态或者各种相似的机器。这样,会使综合征兆有所变化或出现差异。根据差异量的不同,各种综合征兆可被认为是相似的或个别的。在我们的方法中,运用神经网络算法(即 V. Classifier training In order to have a more comprehensive understanding of the working conditions of the machine, more representative samples are needed. These samples can be taken from different machine states or from various similar machines. In this way, there will be some changes in the composite symptoms or differences. Depending on the amount of variability, the various syndromes can be considered similar or individual. In our method, neural network algorithm (ie