Free cookie consent management tool by TermsFeed Policy Generator

Changes between Version 3 and Version 4 of AdditionalMaterial/ECML-PKDD


Ignore:
Timestamp:
06/28/12 12:43:26 (12 years ago)
Author:
gkronber
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • AdditionalMaterial/ECML-PKDD

    v3 v4  
    2424
    2525The following equation shows a non-linear model for the tower data set as identified by symbolic regression in HeuristicLab.
    26 [[Image(model.png)]]
     26[[Image(model_math.png, with=300)]]
    2727
    2828In HeuristicLab a number of different charts and error metrics are available directly in the GUI for each produced solutions. All results are updated dynamically while the algorithm is running.
    29 [[Image(symbolicregressionmodel.png)]]
     29[[Image(screenshots.png)]]
    3030
    3131== Identification of Relevant Variables ==
    3232Frequently it is not necessary to learn a full model of the functional relationship but instead only find a set of relevant variables for the process. This can be achieved easily with HeuristicLab through analysis of relative variable frequencies in the population of models. The following Figure shows a variable frequency chart that clearly shows the six most relevant variables. Notably, the relevance of variables is determined based on non-linear models. So, non-linear influence factors and pair-wise interacting factors can be identified as well. 
    33 [[Image(relevantvariables.png)]]
     33
     34[[Image(relevantvariables_small.png)]]
    3435
    3536== Simplification of Models with Visual Hints ==
    3637A unique feature of HeuristicLab are visual hints for model simplification. By means of visual hints it is very easy to manually prune a complex model as shown above to find a good balance between complexity and accuracy.  The following Figure shows the GUI for model simplification. Green model fragments have a strong impact on the model output while white fragments can be pruned with minimal losses in accuracy, in contrast red fragments increase the error of the model and should be removed. The GUI for model simplification immediately recalculates all error metrics and updates charts dynamically whenever a part of the model is changed by the user.
    37 [[Image(simplification.png)]]
     38
     39[[Image(simplification.png, width=600)]]