Nonlinear regression by the Gauss-Newton method

Question

It is required to implement a nonlinear regression of the circular point cloud.

There is a point cloud in 3d circular cross-section, you need to build a nonlinear regression and find shape defects on the point cloud.

Tell me, in which direction to look for a solution and are there any examples of implementations of nonlinear regression?

2

python машинное-обучение scikit-learn регрессия

Author: 0xdb, 2020-11-16

Source

1 answers

score 4 · Accepted Answer

The first thing that comes to mind for building nonlinear regression models is:

SVR (with RBF and polynomial kernels)
RandomForestRegressor based on a forest / ensemble of decision trees and other more advanced ensembles based on decision trees-such as: LightGBM and CatBoost
neural networks with a linear output layer (a neural network can not be called a linear regression, even though the output layer is linear)

PS examples of using SVR and RandomForestRegressor can be found in the links above