Research code for nonparametric stationarity testing of diffusion processes using time-domain and state-domain volatility estimation. Includes simulation studies, statistical inference, and empirical applications.

This project focuses on forecasting US inflation trends using historical data. Since economic data is often erratic, this notebook explores the transformation of non-stationary series into stationary ones to build reliable predictive models.

End-to-end time series analysis of hourly electricity consumption across 4,400+ London households using smart meter data