I have the following call to numpy.linalg.lstsq:
http://docs.scipy.org/doc/numpy/reference/generated开发者_开发百科/numpy.linalg.lstsq.html
x = [[ 0.69314718] [ 1.09861229] [ 1.38629436] [ 1.60943791] [ 1.79175947] [ 1.94591015]]
y = [ 0. 0.20273255 0.5815754 0.7520387 0.96885669 1.09861229]
l = numpy.linalg.lstsq(x, y)
which returns
l -> tuple: (array([ 0.46323573]), array([ 0.25872885]), 1, array([ 3.63269497]))
Could somebody point out the equivalent function (if available) in
http://commons.apache.org/math/
(or possibly in some other Java math library ...)
Thank you for the pointer Joe.
Here is the code for reference:
double[][] testSquare = {{0.69314718}, {1.09861229}, {1.38629436}, {1.60943791}, {1.79175947}, {1.94591015}};
RealMatrix matrix = MatrixUtils.createRealMatrix(testSquare);
SingularValueDecomposition svd = new SingularValueDecomposition(matrix);
DecompositionSolver ds=svd.getSolver();
double[] b = {0.0, 0.20273255, 0.5815754, 0.7520387, 0.96885669, 1.09861229};
ds.solve(b)[0];
精彩评论