Echo State Network (ESN) Data Augmentation

Echo State Network (ESN) synthetic data generation

The provided MATLAB script implements an Echo State Network (ESN) for the purpose of generating synthetic sequential data based on an initial dataset consisting of randomly generated sequences for a specified number of classes and samples. The script is structured to first create this dataset, then initialize and train the ESN, and finally generate synthetic samples that mimic the statistical properties of the original data. Following the data generation, both original and synthetic datasets are used to train and test a decision tree classifier, evaluating the performance in terms of training and testing accuracy. This approach highlights the potential of ESNs in creating synthetic datasets that can be used to enhance machine learning model training, particularly in scenarios where original data may be limited or imbalanced.