Accurate Weather Prediction on Sunny Days Using Back Propagation Algorithm Compared with Artificial Neural Networks

The objective of this study is to implement a data mining approach to predict weather and climatic changes instantly in real time. For this approach, the novel Back Propagation Classifier (BPC) algorithm is chosen and compared with the Artificial Neural Networks (ANN) classifier. The study involved data collection and model training. Two groups were selected, and the Machine Learning (ML) algorithms employed were BPC and ANN. Each group comprised 20 samples. For SPSS calculations, an 80% G-power value and a 95% confidence interval (CI) were utilized. Result: The BPC demonstrated an impressive accuracy rate of 97.88%, surpassing the accuracy rate of the ANN classifier, which stood at 96.89%. The independent sample t-test yielded a statistical significance value of p=0.002 (p<0.05), indicating a meaningful distinction between the two groups. The implementation of back propagation techniques proved to be more effective than conventional classifiers, leading to a higher level of prediction accuracy.