Comparison of Machine Learning Classification Methods for Weather Prediction: A Performance Analysis

Weather classification is crucial in various sectors, including agriculture, transportation, and disaster management. Accurate weather prediction can help mitigate risks and improve decision-making in these fields. However, classifying weather conditions remains challenging due to the complex and dynamic nature of meteorological data. This study aims to compare different machine learning classification methods to determine the most effective model for weather classification. The research employs a structured methodology consisting of seven key steps: literature study, data understanding, exploratory data analysis, data preparation, modeling, evaluation, and hyperparameter tuning. The study used Decision Tree, Random Forest, Support Vector Machine, K-Nearest Neighbors, Gradient Boosting, AdaBoost, and Extra Trees to identify the best-performing classifier. Model evaluation was conducted using accuracy, precision, recall, and F1-score. The results indicate that Gradient Boosting achieved the highest performance, surpassing other models with an accuracy of 90.15%. To optimize the model further, hyperparameter tuning was conducted using GridSearchCV, and feature selection was done using SelectKBest. This process resulted in an improved accuracy of 90.22%, demonstrating the effectiveness of model optimization.