I am attempting to create a regression model that best describes my data. Due to it's nature, I'd like to fit a quadratic polynomial, using robust loeass regression. In matlab, I'd run the following command to achieve this, where 0.1 is the bandwidth:
The closest thing I've found to this in Java land is the apache math3 library, as shown in this post . I can use this library as follows:
final LoessInterpolator rloess = new LoessInterpolator(0.1, LoessInterpolator.DEFAULT_ROBUSTNESS_ITERS)
However, LoessInterpolator seems to only perform linear regression, which I do not desire. Is it possible to "enable" the quadratic model in the apache math3 library?