Underwater acoustic signal is one of its most vital features in oceanography for predicting the local variants, helps to process the signal, and also assists to discover non-stationary signals. wavelet neural network (WNN) architecture acquires a huge benefit from the artificial neural network (ANN) and time-frequency localizing features of wavelet analysis (WA). However, the system could train using the ANN model, it has tendency of falling into local minima and generating convergence. A wavelet-aided neural network is integrated with a meta-heuristic optimizer named fruit fly algorithm (FFA) that aids our proposed system in which reduces the computational complexity and improves performance measures. The main objective of the proposed work is of modelling an optimized wavelet neural network for efficient measurement of the mean absolute error and the root mean square error. The simulation of the sound signals has been performed with the use of optimized neural networks that has predicted the unknown signal. The proposed model is assessed by comparing the outcomes with the complements of it, i.e. the traditional neural networks.
All Science Journal Classification (ASJC) codes
- Computer Science (miscellaneous)
- Renewable Energy, Sustainability and the Environment