基于时间序列神经网络预测模型的职工出勤记录数据校正方法

Employee attendance record data correction method based on time series neural network prediction model

为提高职工出勤记录数据的正确率,研究基于时间序列神经网络预测模型的职工出勤记录数据校正方法,采用显著误差检测原理设置出勤记录数据未知变量的约束函数,并基于时间序列神经网络预测模型获取校正最优解;以全覆盖视域设计校正流程,实现对职工出勤记录数据的校正。实验结果表明,文章方法能够将错误数据校正为真实数据,具有较高的准确率和效率。

In order to improve the accuracy of employee att endance record data,the correction method of employee attendance record data based on time series neural network prediction model is studied.The principle of significant error detection is used to set the constraint function of un known variables of attendance record data.The optimal correction solution is obtained based on the time series neural network prediction model.The calibration process is designed with full coverage view to realize the correction of employee at tendance record data.The experimental results show that this method can correct the error data into the real data,and has high accuracy and efficiency.