If you can measure it, consider it predicted

Time Series Forecasting for the M5 Competition

This project focuses on the classification of a subjects sleep stage based on their Apple Watch data

A collection of Machine Learning models that detect if a star system contains exoplanets.

My Thesis: Auto-ML Tool for IoT Applications Version 1.0

心跳信号分类预测

Automated feature engineering

Personal research project seeking to evaluate the efficacy of advanced neural architectures for the purpose of estimating energy expenditure from physical activity.

Sector based classification with feature engineering and tsfresh. Looking 3 months momentum of stocks.

Collection of modern tools and machine learning techniques for data analysis and application in some exercises.

Leakage-safe feature engineering, decision tree–based clustering, interpretable rule extraction, and rigorous multi-stage validation (backtesting and walk-forward analysis). The workflow systematically reduces raw features into high-quality, production-ready trading patterns, emphasizing explainability, robustness, and out-of-sample reliability.

Unsupervised learning for the detection of patterns in human activity sensor data

A quantitative trading oriented time series analysis framework designed to systematically extract, identify, expand, and statistically validate recurring market and trade setup patterns from financial price series, returns, indicators, and derived trading signals.

Detecting potential corruption events from public expenditure time-series data

Implementing segmentation and churn prediction with time series of transaction data.

tsfresh is a library that automatically extracts features.