A machine learning ensemble approach to predicting factors affecting the intention and usage behavior towards online groceries applications in the Philippines

Ma Janice J Gumasing; Ardvin Kester S Ong; Madeline Anne Patrice C Sy; Yogi Tri Prasetyo; Satria Fadil Persada

doi:10.1016/j.heliyon.2023.e20644

A machine learning ensemble approach to predicting factors affecting the intention and usage behavior towards online groceries applications in the Philippines

Heliyon. 2023 Oct 4;9(10):e20644. doi: 10.1016/j.heliyon.2023.e20644. eCollection 2023 Oct.

Authors

Ma Janice J Gumasing¹, Ardvin Kester S Ong^{1

2}, Madeline Anne Patrice C Sy¹, Yogi Tri Prasetyo³, Satria Fadil Persada⁴

Affiliations

¹ School of Industrial Engineering and Engineering Management, Mapúa University, Philippines. 658 Muralla St., Intramuros, Manila, 1002, Philippines.
² E.T. Yuchengo School of Business, Mapúa University. 1191 Pablo Ocampo Sr. Ext., Makati, Metro Manila 1205, Philippines.
³ Department of Industrial Engineering and Management, Yuan Ze University, 135 Yuan-Tung Rd., Chung-Li, 32003, Taiwan.
⁴ Entrepreneurship Department, BINUS Business School Undergraduate Program, Bina Nusantara University, Jakarta 11480, Indonesia.

Abstract

The emergence of e-commerce platforms, especially online grocery shopping, is heightened by the COVID-19 pandemic. Filipino consumers started to adapt online due to the strict quarantine implementations in the country. This study intended to predict and evaluate factors influencing the intention and usage behavior towards online groceries incorporating the integrated Protection Motivation Theory and an extended Unified Theory of Acceptance and Use of Technology applying machine learning ensemble. A total of 373 Filipino consumers of online groceries responded to the survey and evaluated factors under the integrated framework. Artificial Neural Network that is 96.63 % accurate with aligned with the result of the Random Forest Classifier (96 % accuracy with 0.00 standard deviation) having Perceived Benefits as the most significant factor followed by Perceived Vulnerability, Behavioral Intention, Performance Expectancy, and Perceived. These factors will lead to very high usage of online grocery applications. It was established that machine learning algorithms can be used in predicting consumer behavior. These findings may be applied and extended to serve as a framework for government agencies and grocers to market convenient and safe grocery shopping globally.

Keywords: Artificial neural network; Online groceries; PMT; Random forest classifier; UTAUT2.