Pit stops are a fundamental and decisive component of Formula-1 race strategy. This paper aims to develop and evaluate a robust machine learning framework capable of predicting the lap on which a driver will make a pit stop. By analyzing historical lap-by-lap data from the 2022–2025 seasons, this study implements a rigorous methodology to predict pit stop occurrences as a binary classification problem. The data preparation phase involved cleaning anomalous data, such as wet-weather races, and engineering critical strategic features like TyreLife and TimeDiffAhead. Three machine learning algorithms: Random Forest (RF), Support Vector Machine (SVM), and Gradient Boosting (XGBoost) are trained and optimized using grid search with five-fold cross-validation. Model performance is evaluated using the F1-score to address the pronounced class imbalance inherent in pit stop events. Analytical findings show that the XGBoost model is superior, achieving an exceptional F1-Score and ROC-AUC. The research concludes that machine learning algorithms, specifically gradient boosting, can successfully model the complex dynamics of Formula 1 strategy with a high degree of accuracy. ML models serve as a valuable decision-support tool, and the feature importance analysis confirms that tire age and recent pace are the most dominant predictors.
The rapid development of the fintech sector has facilitated the emergence of digital multi-level marketing (MLM) schemes, raising concerns about investor protection. Despite extensive literature on MLM schemes and pyramid schemes, there remains a significant research gap regarding the psychological mechanisms and cognitive biases that drive investor participation behavior. This study investigates factors influencing Vietnamese female investors’ intention to participate in fintech MLM schemes, using Pi Network as a case study. Grounded in behavioral finance theories (Prospect Theory and Social Comparison Theory), the model empirically examines the impacts of herding bias and overconfidence bias, explaining participation intention through the mediating effect of the fear of missing out (FOMO) and perceived risk. A quantitative approach was employed using PLS-SEM analysis, with data collected from 264 female investors in Ho Chi Minh City. The results reveal that herding behavior and overconfidence significantly shape investors’ FOMO and perceived risk, with these biases significantly increasing FOMO and decreasing perceived risk. More importantly, these biases, mediated by FOMO and perceived risk, significantly shape participation intention in fintech MLM schemes. This study contributes empirical evidence showing the interaction between high social connectivity and cognitive-bias-driven vulnerabilities in a rapidly expanding and unregulated digital market such as Vietnam. This study has practical implications for policymakers and financial educators in protecting investors from financial schemes by monitoring social media to debunk “safety in numbers” narratives and prioritize the awareness of biases in financial education to mitigate impulse investments.