Opened 8 months ago

Last modified 7 months ago

#3044 accepted enhancement

Add scaling of input features

Reported by: djoedick Owned by: djoedick
Priority: medium Milestone: HeuristicLab 3.3.17
Component: Problems.DataAnalysis Version: branch
Keywords: Cc:

Description

Investigation of ticket #3043 has shown, that scaling of input variables often leads to an improvement in the quality of the results, especially when using a standard deviation. This feature should be implemented as a standard option for upcoming versions.

Change History (7)

comment:1 Changed 8 months ago by mkommend

  • Component changed from ### Undefined ### to Problems.DataAnalysis

comment:2 Changed 8 months ago by djoedick

  • Status changed from new to accepted

comment:3 Changed 8 months ago by djoedick

r17387: Created new branch for SymRegScaling.
r17388: Added parameter for input variable scaling for regression problems.

comment:4 Changed 8 months ago by mkommend

r17389: Allowed modification of column within transformations.

comment:5 Changed 8 months ago by djoedick

r17390: Added new implementation for transformation of input variables in KernelRidgeRegressionModel and added helper methods in transformation base class.

comment:6 Changed 8 months ago by djoedick

r17391: Added transformation of input variables to problem data and created scaled dataset.

comment:7 Changed 8 months ago by djoedick

Current status: Implementation of a transformation of input variables, creation of scaled dataset in symbolic regression problems using the parameter "Scale Inputs"

Further steps: Use a new dataset for modeling.

Last edited 7 months ago by gkronber (previous) (diff)
Note: See TracTickets for help on using tickets.