机器学习 -- 多元回归实现实例

来源：互联网发布：java多线程日志输出编辑：程序博客网时间：2024/06/06 02:15

实现依然基于accord.net framework

1. 二维数组作为输入和一元输出数组
2. 执行回归，得到回归的方差和，以及回归方程（ax+by+c）的各个系数

实现代码

public class MultiRegressionDemo    {        public static void Execute()        {            // 2 dimensions            MultipleLinearRegression target = new MultipleLinearRegression(2, true);            double[][] inputs =            {                new double[] { 1, 8 },                new double[] { 2, 29 },                new double[] { 3, 2 },                new double[] { 0, 1 },                new double[] { 5, 6 },                new double[] { 4, 7 },                new double[] { 14, 17 },            };            double[] outputs = { 9, 4, 3, 2, 5, 12, 7 };            double error = target.Regress(inputs, outputs);            Console.WriteLine("the regression error is '{0}'", error);            // As result, we will be given the following:            double a = target.Coefficients[0]; // a = 0            double b = target.Coefficients[1]; // b = 0            double c = target.Coefficients[2]; // c = 1            Console.WriteLine("line functions is :{0}x + {1}y + {2}", a, b, c);        }    }class Program    {        static void Main(string[] args)        {            MultiRegressionDemo.Execute();            Console.ReadLine();        }    }

运行结果：

2 0