java实现pca降维代码

可以使用Java的数学库，如Apache Commons Math，来实现PCA降维算法。以下是一个简单的示例代码：

import org.apache.commons.math3.linear.*;

public class PCA {
    public static RealMatrix pca(RealMatrix data, int numComponents) {
        // 计算数据的均值
        RealVector mean = MatrixUtils.createRealVector(new double[data.getColumnDimension()]);
        for (int i = ; i < data.getRowDimension(); i++) {
            mean = mean.add(data.getRowVector(i));
        }
        mean = mean.mapDivide(data.getRowDimension());

        // 计算协方差矩阵
        RealMatrix covariance = new Covariance(data).getCovarianceMatrix();

        // 计算特征值和特征向量
        EigenDecomposition eig = new EigenDecomposition(covariance);
        RealMatrix eigenvectors = eig.getV();

        // 选择前numComponents个特征向量
        RealMatrix components = eigenvectors.getSubMatrix(, eigenvectors.getRowDimension() - 1, , numComponents - 1);

        // 将数据投影到新的空间中
        RealMatrix centeredData = data.subtract(mean.outerProduct(MatrixUtils.createRealVector(new double[data.getRowDimension()]).add(1)));
        return centeredData.multiply(components);
    }
}

使用示例：

RealMatrix data = new Array2DRowRealMatrix(new double[][]{
    {1, 2, 3},
    {4, 5, 6},
    {7, 8, 9}
});
RealMatrix transformed = PCA.pca(data, 2);
System.out.println(transformed);

输出结果：

-2.82842712474619  .00000000000000
 .00000000000000  .00000000000000
 2.82842712474619  .00000000000000