基于逻辑回归的代发工资数据差异核对的数学建模

Mathematical Modeling of Difference Checking of Wage Payment Data Based on Logistic Regression

为了高精度核对代发工资数据差异,保证代发工资数据不出现异常,构建了基于逻辑回归的代发工资数据差异核对数学模型。使用基于信息熵聚类的代发工资数据聚类方法,准确分类已代发与未代发工资数据,缩小核对范围;针对分类后获得的已代发工资数据,通过逻辑函数Sigmoid函数实现已代发工资中异常数据的分类,基于逻辑回归构建数据差异核对的数学模型,完成已代发工资数据与实际需代发工资数据差异核对。实验测试中,所建立模型对多家、多类型企业的代发工资差异数据核对错误数为0家,核对精度高,符合银行代发工资数据差异核对要求;在核对代发工资数据差异时,核对耗时不受代发工资数据量影响,核对耗时均为5min。

In order to check the difference of payroll data with high precision and ensure that there is no abnormality,a mathematical model based on logistic regression is constructed.The model uses the information entropy clustering method to classify the paid and unpaid wage data accurately and narrow the check range.For the salary data obtained after classification,the abnormal data in the paid wages can be classified by the logic function sigmoid function.Based on logistic regression,the mathematical model of data difference checking is constructed to check the difference between the paid wage data and the actual wage data.In the experimental test,the number of errors in the model is 0,which meets the requirements of bank payroll data.Moreover,when the model checks the difference of payroll data,the checking time is not affected by the amount of payroll data,and the checking time is 5 min.