Using Text Mining and Predictive Analytics for Understanding Factors Affecting Behavior for Healthy Eating

Healthy eating behavior is an integral part of a healthy lifestyle, yet many individuals do not practice it despite the numerous benefits it offers. A clear understanding of the factors that influence consumers’ healthy eating behavior can provide valuable insights into the adoption and non-adoption of healthy eating practices. The present study leverages text mining methods to extract meaningful insights from online textual data regarding factors affecting healthy eating behavior. This study then integrates the Theory of Planned Behavior and Stimulus-Organism-Response theories to form a comprehensive structural model. This model examines the relationships of various factors influencing consumers’ healthy eating behavior using the Partial Least Squares Variance-Based Structural Equation Modeling (PLS-SEM) method. The application of predictive analytics introduces a PLS-based predictive model that identifies the key factors influencing people’s healthy eating behavior. In terms of structural relationships, PLS-SEM reveals that Perceptions and Subjective Norms do not significantly influence healthy eating behavior, while Motivations and Perceived Behavioral Control are found to have a substantial impact on individuals' healthy eating behavior. Finally, the results of PLS Predict demonstrate that the PLS-based predictive model introduced in this study possesses strong predictive power, effectively forecasting future cases. This study provides a robust framework for understanding and predicting healthy eating behaviors, which can be instrumental in designing effective interventions.