University Links: Home Page | Site Map
Covenant University Repository

ExplainEx: An Explainable Artificial Intelligence Framework for Interpreting Predictive Models

Udenwagu, Nnaemeka E. and Azeta, A. A. and Misra, Sanjay and Nwaocha, Vivian O. and Enosegbe, Daniel L. and Sharma, Mayank Mohan (2021) ExplainEx: An Explainable Artificial Intelligence Framework for Interpreting Predictive Models. In: International Conference on Hybrid Intelligent Systems, 2021, Online.

[img] PDF
Download (206kB)


Artificial Intelligence (AI) systems are increasingly dependent on machine learning models which lack interpretability and algorithmic transparency, and hence may not be trusted by its users. The fear of failure in these systems is driving many governments to demand more explanation and accountability. Take, for example, the “Right of Explanation” rule proposed in the European Union in 2019, which gives citizens the right to demand an explanation from AI-based predictions. Explainable Artificial Intelligence (XAI) is an attempt to open up the “black box” and create more explainable systems which create predictive models whose results are easily understandable to humans. This paper describes an explanation model called ExplainEx which automatically generates natural language explanation for predictive models by consuming REST API provided by ExpliClas open-source web service. The classification model consists of four main decision tree algorithms including J48, Random Tree, RepTree and FURIA. The user interface was designed based on Microsoft.Net Framework programming platform. At the background is a software engine automating a seamless interaction between Expliclas API and the trained datasets, to provide natural language explanation to users. Unlike other studies, our proposed model is both a stand-alone and client-server based system capable of providing global explanations for any decision tree classifier. It supports multiple concurrent users in a client-server environment and can apply all four algorithms concurrently on a single dataset, returning both precision score and explanation. It is a ready tool for researchers who have datasets and classifiers prepared for explanation. This work bridges the gap between prediction and explanation, thereby allowing researchers to concentrate on data analysis and building state-of-the-art predictive models.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Explainable Artificial Intelligence Predictive models Machine learning Interpretable machine learning
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science
Depositing User: Mrs Patricia Nwokealisi
Date Deposited: 07 Sep 2021 15:09
Last Modified: 07 Sep 2021 15:09

Actions (login required)

View Item View Item