An SEM-ANN Approach - Guidelines in Information Systems Research
Lai-Ying Leonga b, Teck-Soon Hewc d, Keng‐Boon Ooie f✉, Garry Wei‐Han Tane g, Alex A. Koohangh i
- aUniversiti Tunku Abdul Rahman, Jalan Universiti
- bUniversiti Tunku Abdul Rahman, Jalan Universiti, Bandar Barat, Malaysia
- cUniversity of Malaya
- dUniversity of Malaya, Kuala Lumpur, Malaysia
- eUCSI University
- fUCSI University, Cheras, Kuala Lumpur, Malaysia
- gUCSI University, Kuala Lumpur, Malaysia
- hMiddle Georgia State University
- iMiddle Georgia State University, Macon, Georgia, USA
- ✉Corresponding author
Papers open inside SciRah from the original publisher or repository; SciRah never hosts them.
Abstract
The application of hybrid Partial-Least-Square-Structural-Equation-Modeling-Artificial-Neural-Network in Information Systems (IS) research has surged over the years. Grounded on a systematic literature review from the list of premier and other promising IS journals, we found several concerns and issues. Hitherto, there are no guidelines for IS researchers for the hybrid PLS-SEM-ANN approach. We unlocked the potential of the hybrid PLS-SEM-ANN in providing better insight and understanding for IS researchers. In addition, best practices and recommendations for the adoption of PLS-SEM-ANN are discussed. The study contributes to advancing IS research by conducting a systematic literature review with Biblioshiny apps from the R studio’s Bibliometrix package and then proposing a comprehensive and robust approach to address the duality nature through the linear-nonlinear and compensatory-non-compensatory relationships. We proposed a guideline and suggested the minimum sample size, best practices and recommendations for reporting the results. We discuss the opportunities and prospects of the hybrid approach.