Rainfall Prediction and Real-Time Weather Monitoring in Bangladesh: A Comparative Analysis of Machine Learning Algorithms

Recent advancements in technology, such as cloud computing, cyber-physical systems, and the Internet of Things (IoT), have significantly enhanced the capacity for efficient handling and transfer of large volumes of environmental data. Accurate rainfall prediction and weather monitoring have become increasingly complex due to climate change and variability. This study leverages various classification algorithms—Linear Regression (LR), Decision Tree (DT), K-Nearest Neighbors (KNN), Support Vector Machines (SVM), and Naive Bayes (NB)—for rainfall prediction, alongside multiple sensors for real-time weather monitoring. The monitored parameters include temperature, humidity, rainfall, barometric pressure, carbon monoxide (CO) levels, air quality, dust concentration, ultraviolet radiation (UV), and geographic location, with data managed via the ThingSpeak cloud platform. Utilizing a dataset from the Bangladesh Meteorological Department (BMD), spanning from 1970 to 2016, the study evaluates the performance of these classification algorithms based on precision, recall, F1-score, and accuracy. The DT algorithm achieved the highest accuracy at 0.8688 and F1-score 0.9226, while the SVM algorithm recorded the lowest accuracy at 0.5381 and F1-score 0.6705. Notably, all models exhibited higher F1 scores relative to their accuracy metrics. This integrated weather monitoring and rainfall prediction system offers valuable potential for accurate environmental assessments in related projects.