Detection of an Asthma attack in an environment Datasets

AbstractAsthma is a chronic breathing problem impacting around 300 million individuals worldwide, with forecasts indicating this figure may increase to around 400 million by 2025. The intricate nature of asthma, influenced by various environmental factors like allergens, pollution, and climatic changes, requires sophisticated forecasting methodologies. This work introduces a novel expert system utilizing Artificial Intelligence (AI) and Machine Learning (ML) methodologies to evaluate and forecast the likelihood of asthma attacks based on real-time environmental factors and personal health information. The research employs sophisticated ML techniques, specifically XGBoost and LightGradient Boosting Machine (lightGBM), to analyze complex datasets comprising patient symptoms and environmental triggers. By integrating advancedalgorithms with rule-based expert knowledge, the system classifies asthma attack risks into high, medium, and low categories. Principal Component Analysis (PCA) was utilized to streamline feature selection, ensuring the most relevant factors are considered while maintaining predictive accuracy.The study's findings demonstrated remarkable predictive performance. XGBoost achieved accuracies of 92% for city day, 93% for city hour, and 93% for patient datasets. LightGBM showed similarly impressive results, with 91% accuracy for city day, 90% for city hour, and a remarkable 97% for patient-specific data. The ensemble model, combining XGBoost and LightGBM, emerged as the most outstanding approach, delivering 99% accuracy for city datasets and a perfect 100% accuracy for patient data.Future developments aim to enhance the system through advanced AI models, improved personalization, real-time data processing, wearable device integration, and potential expansion to predict risks for other respiratory conditions. By continually refining ML algorithms and incorporating comprehensive patient data, this expert system promises to revolutionize asthma management, offering personalized, timely, and accurate predictions that can significantly improve patients' quality of life.Keywords:Asthma Predictions, Machine Learning, Real-time Monitoring, Extreme Gradient Boosting(XGBoost), Light Gradient Boosting Machine (light GBM).