Prediction of Less Error for Water Renewable Energy Generation Using Long Short Term Memory by Comparing CNN Algorithm to Reduce Mean Square Error

This study aims to predict water renewable energy generation using two machine learning algorithms: Long Short Term Memory (LSTM) and Convolutional Neural Networks (CNN). With a sample size of 116 and a power value of 80%, the research analyzed energy data from various sources, applying a threshold of 0.05% and a 95% confidence interval. The CNN algorithm achieved a performance rate of 28.18%, but the study found that the LSTM algorithm significantly outperforms CNN in this context. Specifically, LSTM achieved a much lower error rate of 12.22%, indicating superior performance in predicting water energy generation. Statistical analysis revealed a significant difference with a p-value of 0.042 (p ¡ 0.05) at a 95% confidence level. Therefore, the research concludes that the LSTM algorithm is more effective and accurate compared to CNN for this application.