对于本项目来说,共有1350组公司财务状况数据,每组数据的输入为10维,代表上述的10个指标,输出为1维,代表公司财务状况,为1时表示财务状况良好,为-1时表示财务状况出现问题。从中随机选取1000组数据作为训练数据,350组数据作为测试数据。根据数据维

数,采用的BP神经网络结构为10—6—1,共训练生成10个BP神经网络弱分类器,最后用10

个弱分类器组成强分类器对公司财务状况进行分类。-For this project, a total of 1350 group company' s financial position data, each data input is 10-dimensional, on behalf of the above 10 indicators, output is 1-dimensional, on behalf of the Company' s financial position, financial condition and 1 is for- 1 indicates that the financial position of the problems. 1000 data from randomly selected as the training data, 350 sets of data as test data. According to the data dimension, using BP neural network structure is 10-6-1, trained a total of 10 BP neural network to generate weak classifiers, and finally with 10 weak classifiers strong classifier to classify the company' s financial condition.