Visual data mining: Using self-organizing maps for electricity distribution regulation

Abstract

The electricity distribution regulation and efficiency benchmarking practice in Finland has drawn attention because of its controversial regulatory scheme and arguably efficient electricity distribution sector. This study uses a computational intelligence tool, i.e., Self-Organizing Map (SOM), in the context of electricity distribution efficiency performance visualization. A SOMmodel has been built based on collected data for 2001-2004. It allows the reader to discriminate between the Finnish DSOs' differing operating circumstances. Through clustering and visualization, an overall perspective of the efficiency performance of the DSOs in 2001-2004 is rendered. In addition, such a visualization approach connects the DSO's efficiency performance to its respective operating characteristics, which is otherwise not straightforwardly indicated by only studying efficiency scores. This application provides evidence that visual data mining with the SOM as a complementary approach in electricity distribution regulation and efficiency benchmarking has the potential to be expended to other regulatory practices (e.g., yardstick regulation). © 2011 Springer-Verlag.