CAJ | 학술논문

数据丢失常常会造成损失，需要根据已有数据估计所丢失的数据。本文利用RBF神经网络估计丢失的话费数据，取得了较好的效果。在利用RBF神经网络处理数值估计问题时，要从散布常数、神经元个数和均方误差三个方面归纳最优估计值选择原则。
수거주실상상회조성손실，수요근거이유수거고계소주실적수거。본문이용RBF신경망락고계주실적화비수거，취득료교호적효과。재이용RBF신경망락처리수치고계문제시，요종산포상수、신경원개수화균방오차삼개방면귀납최우고계치선택원칙。
Due to the expense of the lost data,it is important to estimate the lost data according to the existing data. In this paper,the lost telephone charging data are estimated by RBF neural network,which achieves better effect. When RBF neural network is applied to numerical value estimating,it is important to conclude principles of choosing the best estimate value form aspects of the spread constant,nerve cells number and mean squared error.