期刊名称：International Journal of Computer Science & Information Technology (IJCSIT)
印刷版ISSN：0975-4660
电子版ISSN：0975-3826
出版年度：2014
卷号：6
期号：2
页码：87
出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：This paper describes a generic framework for explaining the prediction of probabilistic machine learningalgorithms using cases. The framework consists of two components: a similarity metric between cases thatis defined relative to a probability model and an novel case-based approach to justifying the probabilisticprediction by estimating the prediction error using case-based reasoning. As basis for deriving similaritymetrics, we define similarity in terms of the principle of interchangeability that two cases are consideredsimilar or identical if two probability distributions, derived from excluding either one or the other case in thecase base, are identical. Lastly, we show the applicability of the proposed approach by deriving a metric forlinear regression, and apply the proposed approach for explaining predictions of the energy performance ofhouseholds.
