Evaluate and Compare Machine Learning Models for Temperature Forecasting of Meteorological Data in Hohhot

Air temperature is closely related to daily life, and accurate temperature forecast can improve efficient production and convenient living. Currently, in terms of temperature forecast by machine learning, academic research is mostly based on the performance of a single or a few models on the same dataset, and there is a slight lack of research on comparing the performance of multiple models under the same conditions. This article uses real-world datasets to evaluate and compare 19 popular machine learning models based on 7 evaluation metrics under the same conditions, which helps with in-depth exploration, analysis, and improvement of machine learning models. The experiments show that CatBoost achieved the best performance on all metrics except for the training time.