Air Quality Prediction Using Genetic Algorithm-Based Feature Selection and Machine Learning Techniques for Sustainable Environmental Management

Abstract This paper presents a systematic approach to improving the accuracy of Air Quality Index (AQI) classification by integrating Genetic Algorithm (GA)-based feature selection with machine learning techniques. Initially, a dataset was collected from Kaggle and preprocessing steps applied. Later, Traditional machine learning models, including SVM, Stochastic Gradient Descent (SGD) Classifier, Logistic Regression, Decision Tree, and Naive Bayes, are applied to establish baseline performance. Later, the GA implemented using the DEAP library, is employed to identify the most relevant features, with a Decision Tree Classifier serving as the fitness evaluator. The selected features are used to retrain the of eliminating redundant selection in reducing dimensionality and enhancing machine learning models, resulting in improved performance. This work highlights the effectiveness of GA-based feature selection in reducing dimensionality by eliminating redundant and irrelevant features, thereby enhancing the predictive performance of AQI classification models.