Alternate Support Vector Machine Decision Trees for Power Systems Rule Extractions

Increasing renewable energy penetrations bring complex feasibility and stability problems. Data-driven methods are applied in extracting and embedding these feasibility and stability rules in power system operations and planning. This paper presents a method of alternate support vector machine decision trees for rule extraction problems. The method significantly improves the classical decision-tree-based algorithms' efficiency, stability, and versatility. Finally, we apply the method to several power and energy system scenarios to show its effectiveness.

Automated summary

This paper proposes an alternative support vector machine decision tree method for rule extraction in order to deal with feasibility and stability issues. The method greatly enhances the efficiency, stability, and versatility of traditional decision tree algorithms, and demonstrates its effectiveness in various power and energy system scenarios.