Optimizing predictability of rating scales by differential evolution for the use by collective intelligent information and database systems

Issue title: Special Section: Collective intelligence in information systems

Guest editors: Ngoc Thanh Nguyen, Edward Szczerbicki, Bogdan Trawiński and Van Du Nguyen

Article type: Research Article

Authors: Koczkodaj, Waldemar W.^{a; *; **} | Kakiashvili, T.^b | Li, Feng^c | Wolny-Dominiak, Alicja^d | Masiak, Jolanta^e

Affiliations: [a] Computer Science, Laurentian University, Sudbury, Ontario P3E 2C6, Canada | [b] Sudbury Therapy, Sudbury, Ontario P3E 4M5, Canada, 885 Regent St., Sudbury, Ontario, Canada | [c] Computer Science, Laurentian University, Sudbury, Ontario P3E 2C6, Canada | [d] Department of Statistical and Mathematical Methods in Economics, University of Economics in Katowice, ul. 1 Maja 50, 40–287 Katowice, Poland | [e] Independent Neurophysiological Unit, Department of Psychiatry, Medical University of Lublin, Lublin, Poland

Correspondence: [**] Corresponding author. Waldemar W. Koczkodaj, Computer Science, Laurentian University, Sudbury, Ontario P3E 2C6, Canada. E-mail: wkoczkodaj@cs.laurentian.ca.

Note: [*] Euro Grant: Human Capital.

Abstract:  In this study, differential evolution (DE) optimization is proposed for rating scale predictability improvement. An arbitrary assignment of equal values for rating scale items is used as the classifier although domain experts are aware that the contribution of individual items may vary. Most academic examinations are conducted by the use of rating scales. Rating scales are also used in psychiatry (especially for screening). This study demonstrates that the differential evolution is effective for optimizing the predictability of rating scales.

Keywords: Rating scale, DE

DOI: 10.3233/JIFS-179361

Journal: Journal of Intelligent & Fuzzy Systems, vol. 37, no. 6, pp. 7545-7553, 2019