Predict Influence of Rare Events in Power of Combined Cycle Power Plant by Copula Method

This study implements the “Copula” statistical framework to analyze complex dependencies. We examined the “CCPP” dataset, comprising 9,568 observations from a Combined Cycle Power Plant operating at maximum capacity between 2006 and 2011. Sourced from the UCI Machine Learning Repository, the dataset includes metrics for Temperature (AT), Ambient Pressure (AP), Relative Humidity (RH), Exhaust Vacuum (V), and electrical output (EP). While Copula methods are predominantly utilized in financial sectors for tail dependence and warranty analysis, they offer distinct advantages in modeling rare events. By constructing a linear regression and a Copula-based simulation (“SimuCCPP”), we compared the datasets to assess how outlier events impact predictive modeling’s.