基于多元线性回归、偏最小二乘、神经网络、卡尔漫滤波、径向基网络、主成分分析等等的程序。可用于建模和预测。-Based on multiple linear regression, partial least squares, neural networks, Carl diffuse filtering, radial basis networks, and so on principal component analysis procedure. Can be used for modeling and prediction.