Special Lecture

Date & Time: November 5, 2026 (Day 2), 16:15–17:00
Location: 5F-VEE02, Millennium Hotel Taichung (Onsite)

Professor Junzo Watada, Shimonoseki City University, Japan

Junzo Watada, Ph.D.
Specially Appointed Professor, Faculty of Data Sciences, Shimonoseki City University, Japan; Professor Emeritus at Waseda University, Japan; Life Senior Member of IEEE

Faculty profile → · Full biography →

Title: A Fuzzy Multivariate Approach to Human-Centric Computing and Its Application to Software Engineering

Abstract

Human-centric computing (HCC) emphasizes the role of human perception, judgment, and experience in the analysis and design of information systems. However, human-centered data obtained through questionnaires and subjective evaluations are inherently vague and uncertain, making conventional statistical approaches difficult to apply directly. This talk presents a fuzzy multivariate modeling framework for analyzing such human-centered information and demonstrates its application to software engineering.

The proposed framework consists of three complementary fuzzy multivariate models: (1) a Regression-Based Fuzzy Quantification model (RBFQ), which quantitatively evaluates the relationships between human perceptions and relevant factors; (2) a Fuzzy Qualitative Discriminant Model (FQDM), which identifies and discriminates influential factors under fuzzy conditions; and (3) a Fuzzy Robust Regression Model (FRRM), which provides robust analysis in the presence of uncertainty and variability in observed data. Together, these models provide a systematic methodology for incorporating human perceptions into quantitative software-quality analysis.

The framework is applied to the analysis of residual software bugs in software systems after product delivery. Human perceptions collected through questionnaires are used to identify factors that influence software quality and the number of remaining bugs. The resulting models provide insights into the factors associated with software maintenance and the duration of post-delivery maintenance activities. Empirical analyses using software development data demonstrate the applicability of the proposed fuzzy multivariate approach in a corporate software engineering environment.

This study illustrates how fuzzy multivariate modeling can bridge human perception and quantitative analysis, providing a practical approach to human-centric software quality assessment and supporting more informed decisions concerning software maintenance and development costs.