Ein Vergleich von Datenanalysemethoden für eine Affective Engineering Methode

DFX 2017: Proceedings of the 28th Symposium Design for X, 4-5 October 2017, Bamburg, Germany

Year: 2017
Editor: Dieter Krause, Kristin Paetzold, Sandro Wartzack
Author: Susan, Gretchen Zöller; Tina, Schröppel; Sandro, Wartzack
Series: DfX
Institution: FAU Erlangen-Nürnberg
Section: User-Centered Design
Page(s): 287-298
ISBN: 978-3-946094-20-3-24

Abstract

Affective Engineering (AE) is an engineering genre that deals with users’ subjective value creation in technical product design. Therein, quantitative instruments to map subjective quality criteria are dominant. ACADE is such an instrument that focuses on a long-term alignment of product design impres-sions to the subjective needs of users. Due to its quantitative backbone, sev-eral mathematical analysis methods seem convenient, whereas their specific benefits and drawbacks are not yet clear in the AE context. Therefore, anal-yses of nonlinear regression, artificial neural networks, fuzzy logic systems and hybrids are examined under the aspect of ACADE applicability. Different quality indicators unveil characteristics which the designer may use to mine their potential for future AE analyses.

Keywords: affective engineering, nonlinear regression, fuzzy logic systems, artificial neural networks