Why are street foods consumed? A machine learning ensemble approach to assess consumption intention of street foods

Abstract

Street food promotes a country's society, culture, and economy. Most studies on street food have focused on food quality, eating habits, and eating motivation. However, despite the evident popularity of street foods among tourists, the intention to consume street foods has been underexplored. This study extended the theory of planned behavior by considering five additional factors to provide a more holistic measurement of the intention to consume street foods. Analyzing these factors using machine learning ensemble techniques, specifically the random forest classifier and artificial neural network. The different algorithms were then subjected to their corresponding optimization processes to determine the classification model's best parameters and optimum output for assessing intention to consume street foods in the Philippines, which yielded high accuracy rates of 93 % and 94.6 %, respectively. Variables such as convenience, hedonic eating value, self-efficacy, and attitude were found to be highly significant. The study collected 80,050 data points from 1,061 valid responses, indicating that consumers' decisions regarding the consumption of street food are influenced by individuals who hold significant importance in their lives. Moreover, consumers exercise their agency in the consumption of street foods, utilizing their resources and patronage to contribute to the societal and economic benefits associated with such consumption within the country. Given the freshness of most street food items being prepared on-site, this instilled greater confidence in their choice. The findings of on the intention to consume street food in a cultural perspective can be leveraged by government officials to promote and advertise street foods, highlighting their connection to the country's culture, diversity, and society.