ecg5000

Implementation of Echo State Networks (ESN) with experiments on MNIST and ECG5000. Includes comparison with Linear Regression and analysis of weight initialization methods for time-series and classification tasks.

A reproducible time-series anomaly detection pipeline comparing Conv1D and LSTM Autoencoders with classical baselines on the ECG5000 heartbeat dataset.

Implementácia metód Isolation Forest, One-Class SVM a Autoencoder na detekciu anomálií pri nedostatku anomálnych dát na datasetoch NASA Battery, NAB a ECG5000.

ECG heartbeat anomaly detection comparing LSTM, Bi-LSTM and XGBoost, with a live Streamlit demo.

Implementation of Isolation Forest, One-Class SVM, and Autoencoder Methods for Anomaly Detection under Limited Anomalous Data Using the NASA Battery, NAB, and ECG5000 Datasets.