Regresar

Analysis of Prediction and Clustering of Electricity Consumption in the Province of Imbabura-Ecuador for the Planning of Energy Resources

Abstract:

The electrical planning of a country is a growing need for its economic development. However, this analysis is complex to carry out due to the different consumption habits of the participants in this market. In this sense, in Imbabura-Ecuador, residential electricity consumption is the one that needs the greatest electricity demand. For this reason, a prediction and grouping analysis by municipalities is presented using machine learning algorithms in order to determine consumer trends and present reports for proper electrical planning. As relevant results, the models of decision support machines and random forests proved to be suitable for this task with a prediction error of less than 10%. For its part, the k-means algorithm was able to group four types of electricity consumption with a representation of 98% of the data variability.