Home // International Journal On Advances in Software, volume 13, numbers 1 and 2, 2020 // View article

An Explanation Framework for Whole Processes of Data Analysis Applications: Concepts and Use Cases

Authors:
Hiroshi Ishikawa
Yukio Yamamoto
Masaharu Hirota
Masaki Endo

Keywords: social big data; explanayion; data model; data management; data mining.

Abstract:
The main contribution of the paper is to address the necessity of both macro and micro explanations for Social Big Data (SBD) applications and to propose an explanation framework integrating both, allowing SBD applications to be more widely accepted and used. The framework provides both a macro explanation of the whole procedure and a micro explanation of the constructed model and an explanation of the decisions made by the model. Application systems including Artificial Intelligence (AI) or Data Mining (DM) need reproducibility to ensure their reliability as scientific systems. For that purpose, it is important to illustrate the procedures of the system explicitly and abstractly (that is, macro explanations). This paper has scientific value in that it proposes a data model for that purpose and illustrates the possibility of macro explanations through one use case of social science. Scientists also need to provide evidence that the results obtained by AI or DM are valid. In other words, this paper also has scientific value in that it reveals how the features of the model and concrete grounds for judgment can be explained through two use cases of natural science.

Pages: 1 to 15

Copyright: Copyright (c) to authors, 2020. Used with permission.

Publication date: June 30, 2020

Published in: journal

ISSN: 1942-2628