CAJ | 학술논문

在研究多个自变量对多个因变量回归分析的数学模型和神经网络的基础上,给出了BP算法求解多元回归模型.采用JAVA程序设计语言,开发了基于Web Service的回归分析网络系统,实现了对回归参数和主要评估指标的预测.实际应用表明,利用该系统得到的拟合值精度较高,预测结果和实测结果吻合较好,可以用于解决预测领域中的多元回归问题.
재연구다개자변량대다개인변량회귀분석적수학모형화신경망락적기출상,급출료BP산법구해다원회귀모형.채용JAVA정서설계어언,개발료기우Web Service적회귀분석망락계통,실현료대회귀삼수화주요평고지표적예측.실제응용표명,이용해계통득도적의합치정도교고,예측결과화실측결과문합교호,가이용우해결예측영역중적다원회귀문제.
Based on research of regression analysis model of multiple independent variable vs. multiple dependent variable and neural network, this present paper put forward a solution to solve multivariate regression model applying BP algorithm. Regressive analysis network system based on Web Service was developed using JAVA language, which implemented the prediction of regression parameters and main evaluation index. By practical application, this network system can obtain high precise fitted value and predicted results, which anastomosed measured results well. So it could be used for the multiple regressive analysis in forecasting area.