Mapping energy sustainability using the Kohonen self-organizing maps - Case study

Abstract

There is no one-size-fits-all solution for global energy sustainability. It is affected by a country's policies, resources, financing opportunities, the attention given to it, and the degree to which other pressing issues, such as poverty and access to power, are addressed in each country. This paper used 28 indicators to assess a country's energy sustainability. These indicators ranged from CO2 per capita emissions to the share of electricity production from wind. The indicators were adjusted so that the lower values were more sustainable. 31 countries were selected. Kohonen's self-organizing maps were used to present this 28-dimensional model in a 2D map. This algorithm is a type of artificial neural network trained by unsupervised learning. Eight clusters of countries were found through the analysis using the four-light traffic panel, where green is the most sustainable, followed by blue, yellow, and red. The closeness on the map does not guarantee that two countries will be in the same cluster. Three color matches and closeness were the criteria for the creation of clusters. The energy transition from the map's lower to upper sides is plainly visible. Red clusters make up the lower side, while green clusters make up the top. One blue cluster and two yellow clusters are located between them. The algorithm converged in the 133rd iteration; the quantization error reached 0 and the topological error reached a value of 25. This method can be used to visualize how far or close a country is to the best achievers. It shows which parameters can or should be improved and what the world trend is. If there is a trend, cost-effective solutions are frequently present.