Neuro-Symbolic AI (for True Epub) by Alexiei Dingli
 David Farrugia

Author:Alexiei Dingli
, David Farrugia
Language: eng
Format: epub
Publisher: Packt Publishing Pvt Ltd
Published: 2023-05-23T00:00:00+00:00

Permutation importance

Permutation importance measures the input feature’s significance based on the increase in the model’s error. The underlying assumption is that the model’s error should also change if we shuffle the values of the pattern. Suppose the change in values does not affect the global error level. In that case, it simply means that the feature is probably unimportant since, irrespective of its value, it has almost no effect on the final output. On the other hand, if a change in the permutations drastically affects the output error, then there is a strong link, and the feature cannot be ignored when devising an explanation for that model. To identify these features, one of them is removed from the dataset, the estimator is re-trained, and the score is checked. Since we’re working on the whole model, permutation importance only works at a global level.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.