Workshop temperature and humidity prediction based on machine learning algorithms

The tobacco leaf processing workshop lacks constant temperature control, causing its production environment to be susceptible to external climatic influences, leading to fluctuations in product quality. To improve product quality stability, this paper uses historical indoor and outdoor temperature and humidity data from February 2023 to August 2024. Feature engineering was employed to enhance data representation, and a Random Forest (RF) model was used for temperature and humidity prediction. The results indicate that, under optimal parameters, the root mean square error (RMSE) for temperature prediction is 0.01, and for humidity prediction is 0.26, demonstrating high predictive accuracy. By accurately forecasting future workshop temperature and humidity, the system provides data support to operators, effectively improving the accuracy of the moisture setting in the tobacco processing process and ensuring product quality.