Hypertension Detection From High-Dimensional Representation of Photoplethysmogram Signals

mlots is python package that provides Machine Learning tools for Time-Series Classification.

Zero-Shot ECG Generalization using Morphology-Rhythm Disentanglement and Mamba State Space Models. Features a production-ready Clinical Dashboard

A production-grade deep learning framework for zero-shot ECG classification that achieves state-of-the-art generalization through morphology-rhythm disentanglement and efficient long-range sequence modeling with Mamba/SSM.

Intra-operative hypotension prediction

Reproduction and empirical evaluation of MiniRocket (KDD 2021) for COMP41850

CUDA implementation of the MiniRocket algorithm

Multimodal time-series pipeline for ERR@HRI 2.0: MiniRocket+Ridge to detect robot errors and user repairs.

EEG Classification using MiniRocket + Ridge Classifier on BCI Competition IV Dataset 2a. Reproducible pipeline for Time-Series feature extraction and classification.

ML pipeline for classifying Tidal Disruption Events from LSST astronomical light curves — including multi-output Gaussian process resampling of time series, using feature extractors like MiniRocket and light_curve, and binary classifiers like LightGBM and logistic regression.

Hard drive failures can lead to data loss and increased bussiness expenses. In this project we detect hard drive failure before it happens with sensor data in a time series format. We transform the data into a linear format with convolutional kernels and classify with a random forest model.

PyTorch implementation of MiniRocket with soft PPV.

Vibration-based predictive maintenance dashboard for rotary machinery — FastAPI + sktime MiniRocket . Student data science consulting project for Machinery Monitoring Systems, LLC × Institute for AI and Data Science, University at Buffalo.

HG4TS_SSF: Hierarchy Generation for Time Series using Stochastic Splitting Functions