An application of data mining algorithms for predicting factors affecting Big Data Analysis adoption readiness in SMEs

Abstract

The adoption of Big Data Analysis (BDA) has become popular among firms since it creates evidence for decision-making by managers. However, the adoption of BDA continues to be poor among small and medium enterprises (SMEs). Therefore, this study adopted the Technology-Organization-Environment (TOE) framework to identify the drivers of readiness to adopt BDA among SMEs. Chi-square automatic interaction detection (CHAID), Bayesian network, neural network, and C5.0 algorithms of data mining were utilized to analyze data collected from 240 Vietnamese managers of SMEs. The evaluation model identified the C5.0 algorithm as the best model, with accurate results for the prediction of factors influencing the readiness to adopt BDA among SMEs. The findings revealed management support, data quality, firm size, data security and cost to be the fundamental factors influencing BDA adoption readiness. Moreover, the results identified the service sector as having a higher level of readiness toward the adoption of BDA compared to the manufacturing sector. The findings are imperative for the enhancement of the decision-making process and advancement of comprehension of the determinants of BDA adoption among SMEs by researchers, managers, providers and policymakers.

Keywords

• bayesian networks
• big data analysis
• chaid
• c5.0
• data mining
• neural network
• smes