Performance Comparison of Flood Prediction Using Recurrent Time Delay Neural Network and K-Nearest Neighbor Algorithm
D. Anjireddy, K. Jaisharma · 2024
Floods have a significant threat to the living things, also it causes extensive damages to the landscape. This research article focuses on experimenting with the flood prediction model using the Novel Recurrent Time Delay Neural Network (NRTDNN) and comparing its performance with the K-Nearest Neighbor (KNN) model based on the accuracy. The study involves two groups of participants: The first group, called NRTDNN, consists of 20 individuals, and the second group, named KNN, has the same number of participants. To perform the calculation, used G-power 0.80 as the software tool. We choose an alpha of 0.05, beta as 0.02, which corresponds to a 95% confidence interval. These parameters allow us to test the statistical hypotheses with high accuracy and reliability. The research results showed that the Novel Recurrent Time Delay Neural Networks algorithm outperformed the K-Nearest Neighbour algorithm in terms of accuracy. The former achieved 91.40% accuracy, while the latter only attained 85.11% accuracy. This indicates that the Novel Recurrent Time Delay Neural Networks algorithm is more suitable for the task at hand than the K-Nearest Neighbor algorithm. To test the difference between the means of the two groups, an independent sample T-Test was performed. The analysis yielded a p-value of 0.000, which deserves as significance of the result. Therefore, the null hypothesis was rejected and the difference was deemed statistically significant. The proposed methodology improves the accuracy of flood prediction, as shown by the study. The Novel Recurrent Time Delay Neural Network outperforms the K-nearest Neighbor (KNB) algorithm in prediction accuracy rate, achieving 91.40% versus 85.11%. Building on previous work, the NRTDNN research improves the accuracy of flood prediction beyond the KNB algorithm. This improvement allows for more effective interventions in the face of climate change consequences.