通过运用主成分分析法，线性分类器，神经网络来对测试样本与训练样本加以匹配-To be matched to the test sample and the training samples by using principal component analysis, linear classifiers, neural network

实现对于数据特征的识别和提取，进而实现重构，重构值与输入数据相差越小越好，对于该深度学习网络的训练可以分为预训练和调优过程，对此需要把数据按比例分为训练集和测试集，进行系数的调整，从而实现数据重构。-For realization identify and extract data features, thus achieving the reconstruction, reconstruction of the input data value difference as small as p