ApeRIodic SEmi Parametric

ApeRIodic SEmi-parametric (ARISE) processes aim to model complex, real-world data exhibiting long-term memory and non-stationarity, particularly in financial markets and time series analysis, without relying on restrictive assumptions like periodicity or Gaussianity. Current research focuses on developing robust and efficient algorithms, including those based on kernel methods, conditional flows, and integral probability metrics, to estimate parameters and quantify uncertainty in these models. This work has implications for various fields, improving the accuracy of probabilistic forecasting, anomaly detection in complex networks, and causal inference by providing more flexible and adaptable modeling approaches.