Machine learning to explore the stochastic perturbations in revenue of pandemic-influenced small businesses

• Published in: Nonlinear dynamics Vol. 112; no. 2; pp. 1549 - 1558
• Main Authors: Yu, Zhenhua, Sohail, Ayesha
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.01.2024; Springer Nature B.V
• Subjects: Automotive Engineering; Classical Mechanics; Control; Dynamical Systems; Engineering; Machine learning; Mathematical models; Mechanical Engineering; Modelling; Original Paper; Pandemics; Perturbation; Randomness; Revenue; Risk assessment; Stochastic processes; Vibration; Algorithm development; COVID-19 pandemic; Revenue; SVM; Clinical practice; Risk assessment
• ISSN: 0924-090X; 1573-269X
• DOI: 10.1007/s11071-023-09011-7

The classical models of risk assessment and the forecasting tools of business revenue typically contain unknown parameters with a range of empirical values that are insufficient for making intuitive predictions. To improve the modeling and simulation approaches, it is necessary to deal with the associated data-sets, subject to several complex constraints, using smart programming tools. These constraints are responsible for the randomness, noise and perturbation, technically termed as the stochastic effects. Such stochastic processes, when incorporated with seasonality lead to the mean reverting L’evy-based Ornstein–Uhlenbeck approach. The Ornstein–Uhlenbeck modeling approach is used here for the assessment of the revenue. Regression learner models of machine learning are developed to explore impact of the change in temperature on pandemic thresholds and with this, the change in revenue. The current research strategy can prove to support the investors in their future investment planning and decisions and to forecast the risks and the fate of small and individual-based businesses.