基于冗余数据压缩算法的经济信用风险研究

Research on economic credit risk of based on redundant data compression algorithm

针对信贷市场中的大数据难以有效分析经济信用风险的问题,本研究以终端云计算存储网络平台的金融数据包为基础,通过筛选过滤和降维处理的操作,在静态的分块模块中将上传的数据包文件进行分割得到不同容量的数据包从而实现对冗余数据的高纬度的立体空间映射;动态的分块模块中时间序列数据副本边缘特征空间向量的迭代操作步骤数对冗余数据包进行模式的转换。选取国民经济中划分的18个行业81个月实验结果表明:时间滑动窗口为5 s,对象个数为40个时,经济信用风险的误差在1%左右,数据误差和标准差分别在3%和4%以内,数据集检出准确率超过95%;冗余数据的压缩率为5.93%以内。

For the credit markets effectively analyze large data difficult economic problems of credit risk, this study terminal cloud storage network platform of financial data packet basis, by operating the screening and filtration reducing the dimension of the block in the module, in a static Upload file packets obtained by dividing the capacity of different packets of three-dimensional space in order to achieve high latitudes redundant data mapping; Iterative procedure number of copies of the data dynamic partitioning module for the time series feature space vector edge of redundant data packet conversion mode. Select the national economy division 18 industry 81 months results show that： the sliding window of time 5 s, the number of objects is 40, the credit risk in the economy error of about 1%, the data error and standard deviation, respectively, and 3% less than 4%, the detection accuracy of the data set more than 95% ; redundant data compression rate of less than 5.93%.