Automatic Analysis Pipeline

Gausian Process Regression

1 Overview

This pipeline utilizes a random forest model to predict properties of groups outside the training set. A validation set is utilized to check the predictions. Results are dependent on the before used filter method in Data_Overview.qmd.

2 Model Creation

2.1 Parameter Testing

Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 
Using automatic sigma estimation (sigest) for RBF or laplace kernel 

3 Predictions

3.1 Compare Predictions

3.1.1 MAPE per Group [%]

Group	σ_m	ε_m	σ_b	ε_b
F21	4.2	4.4	7.8	61.4
F37	9.7	4.2	18.0	139.2
F4	13.5	7.1	20.9	32.0
F48	13.1	40.2	15.4	114.6

3.1.2 R² for all Groups

Group	σ_m	ε_m	σ_b	ε_b
R² (gesamt)	0.013	-0.093	-0.121	-1.692