ML IALGO LINREG: Difference between revisions

Revision as of 18:16, 12 October 2021

ML_IALGO_LINREG = [integer]
Default: ML_IALGO_LINREG = 1

Description: This tag determines how to solve the system of linear equations in the ridge regression method for machine learning.

The following options are available:

ML_IALGO_LINREG=1: Bayesian linear regression (see here). Usable with NSW $\ge$ 1.
ML_IALGO_LINREG=2: QR factorization. Usable with NSW=0,1.
ML_IALGO_LINREG=3: Singular value decomposition. Usable with NSW=0,1.
ML_IALGO_LINREG=4: Singular value decomposition with Tikhonov regularization. Usable with NSW=0,1.

@@ Line 1: / Line 1: @@
 {{TAGDEF|ML_IALGO_LINREG|[integer]|1}}
-Description: This tag determines how to solve the linear equations in the ridge regression for machine learning.
+Description: This tag determines how to solve the system of linear equations in the ridge regression method for machine learning.
 ----
 The following options are available:
-*{{TAG|ML_IALGO_LINREG}}=1: Bayesian linear regression (see [[Machine learning force field: Theory#Bayesian error estimation|here]]. This is the only method usable with on the fly learning ({{TAG|NSW}}>1.
+*{{TAG|ML_IALGO_LINREG}}=1: Bayesian linear regression (see [[Machine learning force field: Theory#Bayesian error estimation|here]]). Usable with {{TAG|NSW}}<math>\ge</math>1.
-*{{TAG|ML_IALGO_LINREG}}=2:
+*{{TAG|ML_IALGO_LINREG}}=2: QR factorization. Usable with {{TAG|NSW}}=0,1.
-*{{TAG|ML_IALGO_LINREG}}=3:
+*{{TAG|ML_IALGO_LINREG}}=3: Singular value decomposition. Usable with {{TAG|NSW}}=0,1.
-*{{TAG|ML_IALGO_LINREG}}=4:
+*{{TAG|ML_IALGO_LINREG}}=4: Singular value decomposition with Tikhonov regularization. Usable with {{TAG|NSW}}=0,1.
 == Related Tags and Sections ==