Abstract
Abstract In this paper the goal is to explain predictions from complex machine learning models. One method that has become very popular during the last few years is Shapley values. The original development of Shapley values for prediction explanation relied on the assumption that the features being described were independent. If the features in reality are dependent this may lead to incorrect explanations. Hence, there have recently been attempts of appropriately modelling/estimating the dependence between the features. Although the previously proposed methods clearly outperform the traditional approach assuming independence, they have their weaknesses. In this paper we propose two new approaches for modelling the dependence between the features. Both approaches are based on vine copulas, which are flexible tools for modelling multivariate non-Gaussian distributions able to characterise a wide range of complex dependencies. The performance of the proposed methods is evaluated on simulated data sets and a real data set. The experiments demonstrate that the vine copula approaches give more accurate approximations to the true Shapley values than their competitors.
Highlights
In many applications complex machine learning models like Gradient Boosting Machines, Random Forest and Deep Neural Networks are outperforming traditional regression models
The original development of Shapley values for prediction explanation relied on the assumption that the features being described were independent
In this paper we propose two new approaches for modelling the dependence between the features. Both approaches are based on vine copulas, which are exible tools for modelling multivariate non-Gaussian distributions able to characterise a wide range of complex dependencies
Summary
In many applications complex machine learning models like Gradient Boosting Machines, Random Forest and Deep Neural Networks are outperforming traditional regression models. Existing work on explaining complex models may be divided into two main categories; global and local explanations The former try to describe the model as whole, in terms of which variables/features in uenced the general model the most. On the other hand, try to identify how the di erent input variables/features in uenced a speci c prediction/output from the model, and are often referred to as individual prediction explanation methods. Such explanations are useful for complex models which behave rather di erent for di erent feature combinations, meaning that the global explanation is not representative for the local behavior.
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
