How to Explain Individual Classification Decisions
Machine Learning
2009-12-08 v1 Machine Learning
Abstract
After building a classifier with modern tools of machine learning we typically have a black box at hand that is able to predict well for unseen data. Thus, we get an answer to the question what is the most likely label of a given unseen data point. However, most methods will provide no answer why the model predicted the particular label for a single instance and what features were most influential for that particular instance. The only method that is currently able to provide such explanations are decision trees. This paper proposes a procedure which (based on a set of assumptions) allows to explain the decisions of any classification method.
Cite
@article{arxiv.0912.1128,
title = {How to Explain Individual Classification Decisions},
author = {David Baehrens and Timon Schroeter and Stefan Harmeling and Motoaki Kawanabe and Katja Hansen and Klaus-Robert Mueller},
journal= {arXiv preprint arXiv:0912.1128},
year = {2009}
}
Comments
31 pages, 14 figures