A NCaRBS analysis of SME intended innovation: Learning about the Don't Knows. (March 2016)

• Record Type: Journal Article
• Authors: Beynon, Malcolm J; Jones, Paul; Pickernell, David; Packham, Gary
• Is Part Of: Omega. Volume 59:Part A(2016)
• Page Start: 97
• Page End: 112
• Publication Date: 2016-03
• Subjects: SME -- NCaRBS -- Don't Know -- Innovation; Management -- Periodicals; 658.4005
• DOI: 10.1016/j.omega.2015.04.018
• Languages: English
• ISSNs: 0305-0483

Abstract

Abstract: This study demonstrates a novel form of business analytics, respecting the quality of the data available (allowing incompleteness in the data set), as well as engaging with the uncertainty in the considered outcome variable (inclusive of Don't Know (DK) responses). The analysis employs the NCaRBS technique, based on the Dempster–Shafer theory of evidence, to investigate the relationship between Small and Medium-sized Enterprise (SME) characteristics and whether they intended to undertake future innovation. The allowed outcome response for intended innovation was either, Yes, No and DK, all of which are considered pertinent responses in this analysis. An additional consequence of the use of the NCaRBS technique is the ability to analyse an incomplete data set, with missing values in the characteristic variables considered, without the need to manage their presence. From a soft computing perspective, this study demonstrates just how exciting the business analytics field of study can be in terms of pushing the bounds of the ability to handle real 'incomplete' business data which has real, and sometimes uncertain, outcomes. Further, the findings also inform how different notions of ignorance in evidence are accounted for in such analysis. Graphical abstract: Highlights: This study demonstrates a novel form of business analytics as well as engaging with the uncertainty in the considered outcome variable. The analysis employs the NCaRBS technique to investigate the relationship between Small and Medium-sized Enterprise (SME) characteristics and whether they intended to undertake future innovation. The NCaRBS technique allows the ability to analyse an incomplete data set, with missing values in the characteristic variables considered, without the need to manage their presence. This study demonstrates how to handle real 'incomplete' business data which has real, and sometimes uncertain, outcomes. The findings also inform how different notions of ignorance in evidence are accounted for in such analysis.