This study uses machine learning approach to detect food safety risks and identify potential solutions by analyzing consumer reviews in the online catering platform.
The approach involves analyzing over 10,000 consumer reviews to develop dictionary of food safety risk-related keywords. These methods allow for the effective identification of specific food safety risks from consumer-generated online reviews. Additionally, a best-worst scaling experiment complemented by mixed logit model analysis ranks potential solutions according to consumer preferences, integrating these insights into practical strategies.
The study evaluates consumer preference to various food safety risk solutions using best-worst scaling experiment, indicating strong preferences for the food safety traceability system and legislative measures as effective strategies. Meanwhile, other strategies such as science popularization and merchants’ elimination evoke mixed reactions, reflecting diverse consumer perceptions on their efficacy. These insights are pivotal for shaping targeted, effective food safety practices in online catering platforms.
While the findings offer significant insights, the research is limited by the specificity of the data to online reviews, which may not fully capture the breadth of food safety risks encountered by consumers. Future research could expand the scope of data sources to include more direct consumer interactions and feedback mechanisms.
This study provides valuable strategies for online catering platforms to integrate the solutions of addressing food safety risks into their business models effectively.
Consumer preferences for online food safety solutions determine their efficacy. Few studies have examined this problem from consumers’ perspectives. This study reminded academics of consumer desire for different solutions. The preceding data analysis showed that respondents expect internet food safety issues can be solved before they occur. Thus, developing food safety solutions before they reach customers is advised.
This research extends the conventional use of text mining for sentiment analysis by applying it within the food safety context in the online catering platform. It uniquely combines the analytical rigor of machine learning with practical marketing strategies to address a critical public health issue.
