Abstract
Background/Objectives: Widely known for its therapeutic benefits, hydrotherapy utilizes water's physical properties, such as temperature, hydrostatic pressure, and viscosity, to influence physiological responses. Among these, body temperature modulation plays a crucial role in enhancing circulatory function, muscle relaxation, and metabolic processes. While hydrotherapy can improve systemic health, particularly cardiac function, improper temperature control poses risks, especially for vulnerable populations such as the elderly or individuals with thermoregulatory impairments. Therefore, accurately predicting post-bath body temperature is essential for ensuring safety and optimizing therapeutic outcomes. Methods: This study explored the use of fuzzy inference systems to predict post-bath body temperature, leveraging an adaptive neuro-fuzzy inference system, evolutionary fuzzy inference system (EVOFIS), and enhanced Takagi-Sugeno fuzzy system. These models were compared with random forest and support vector machine models using hydrotherapy-related data. Results: The results show that EVOFIS outperformed other models, particularly in predicting deep body temperature, which is clinically significant as it is closely linked to core physiological regulation. Conclusions: The ability to accurately forecast deep-temperature dynamics enables proactive management of hyperthermia risk, supporting safer hydrotherapy practices for at-risk groups. These findings highlight the potential of FIS-based models for non-invasive temperature prediction, contributing to enhanced safety and personalization in hydrotherapy applications.