JAVA中的多项式回归

  某天在摸鱼的焱童鞋写，被领导抓取给其他部门做一个c#平台功能的迁移工作，不用1天基本框架都迁移完毕，有余下：c#的回归分析和out参数方法 需要新增

使用gpt了解了可以用commons-math3中的相关方法来模拟c#中mathNet库里的fix方法,java没有out参数，使用相关数组进行转化，步骤如下

1.导入依赖

<dependency>
            <groupId>org.apache.commons</groupId>
            <artifactId>commons-math3</artifactId>
            <version>3.6.1</version>
        </dependency>

2.相关代码

public static Double[] PolyFitAndEvaluate(Double[] x, Double[] y, int degree, Double[] para) {
    WeightedObservedPoints points = new WeightedObservedPoints();
    for (int i = 0; i < x.length; i++) {
        points.add(x[i], y[i]);
    }

    PolynomialCurveFitter fitter = PolynomialCurveFitter.create(degree);
    double[] coefficients = fitter.fit(points.toList());

    // 将 coefficients 数组转换为 Double 类型并复制到 para 数组中
    for (int i = 0; i < coefficients.length; i++) {
        para[i] = coefficients[i];
    }

    // 创建 PolynomialFunction 对象
    PolynomialFunction polynomialFunction = new PolynomialFunction(coefficients);

    // 评估多项式
    Double[] y0 = new Double[x.length];
    for (int i = 0; i < x.length; i++) {
        y0[i] = polynomialFunction.value(x[i]);
    }

    return y0;
}

3.测试

测试数据： 



public static void main(String[] args) {
    Double[] x = { 1d, 2d, 3d };
    Double[] y = { 2d, 4d, 6d };
    int degree = 2;
    Double[] para = new Double[degree + 1];

    Double[] y0 = PolyFitAndEvaluate(x, y, degree, para);
    for (double val : y0) {
        System.out.println(val);
    }
    // 输出 para 数组
    for (double val : para) {
        System.out.println(val);
    }
}