多元线性回归正规方程java代码

正规方程：$A=(XX^T)^{-1}X^TY$ 之前已经证明过了。
用JAMA包做矩阵计算
结果自己造的数据矩阵不可逆。。。。。

package com.zy.ml;

import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;

import org.apache.commons.io.FileUtils;
import org.apache.commons.io.LineIterator;

import Jama.Matrix;

/**
 * 多元线性回归正规方程，最小二乘 A=(X*X^T)^-1*X^T*Y
 * 
 * @author yzhang
 *
 */
public class LinearRegressionLSM {
    static List<String> trainData = new ArrayList<String>();

    static int m = 0;

    public void getTrainDate(String file) {
        LineIterator li = null;

        try {
            li = FileUtils.lineIterator(new File(file));
        } catch (IOException e) {
            e.printStackTrace();
        }
        for (; li.hasNext();) {
            String line = li.nextLine();
            trainData.add(line);
        }
        m = trainData.size();
    }

    private double[][] getX() {
         double[][] x = new double[m][3];
        for (int i = 0; i < m; i++) {
            x[i][0] = 1.0;
            x[i][1] = Double.parseDouble(trainData.get(i).split(",")[0]);
            x[i][2] = Double.parseDouble(trainData.get(i).split(",")[1]);
        }
        return x;
    }

    private double[][] getY() {
         double[][] y = new double[m][1];
        for (int i = 0; i < m; i++) {
            y[i][0] = Double.parseDouble(trainData.get(i).split(",")[2]);
        }
        return y;
    }

    public Matrix getTheta() {
        double[][] x = getX();
        double[][] y = getY();
        Matrix XM = new Matrix(x);
        Matrix YM = new Matrix(y);
        Matrix A = (  ((   XM.times(XM.transpose())   ).inverse() ) .times(XM.transpose()) ).times(YM);
        return A;
    }

    public static void main(String[] args) {
        LinearRegressionLSM lr = new LinearRegressionLSM();
        lr.getTrainDate("/home/test.txt");
        Matrix A = lr.getTheta();
        double[][] array1 = A.getArray();
        for (int i = 0; i < array1.length; i++) {
            for (int j = 0; j < array1[i].length; j++) {
                System.out.print(array1[i][j] + " ");
            }
            System.out.println(" ");
        }

        // double[][] array = { { -1, 1, 0 }, { -4, 3, 0 }, { 1, 0, 2 } };
        // // 定义一个矩阵
        // Matrix A = new Matrix(array);
        // //转置
        // Matrix B = A.transpose();
        // double[][] array1 = B.getArray();
        // for (int i = 0; i < array1.length; i++) {
        // for (int j = 0; j < array1[i].length; j++) {
        // System.out.print(array1[i][j] + " ");
        // }
        // System.out.println(" ");
        // }
        // System.out.println(" -------");
        // //逆矩阵
        // Matrix C = A.inverse();
        // double[][] array2 = C.getArray();
        // for (int i = 0; i < array2.length; i++) {
        // for (int j = 0; j < array2[i].length; j++) {
        // System.out.print(array2[i][j] + " ");
        // }
        // System.out.println(" ");
        // }
        // System.out.println(" -------");
        // Matrix D = A.times(C);
        // double[][] array3 = D.getArray();
        // for (int i = 0; i < array3.length; i++) {
        // for (int j = 0; j < array3[i].length; j++) {
        // System.out.print(array3[i][j] + " ");
        // }
        // System.out.println(" ");
        // }
    }

}