FEATURIZATION OF ENERGY CONSUMPTION CHARACTERISTICS FOR DATA-DRIVEN MANAGEMENT OF MULTIPLE BUILDING (PART 2) : APPLICATION OF THE RE-CLUSTERING PROCESS FOR THE PURPOSE OF CLEAR CATEGORIZATION IN MULTISTAGE ANALYSIS METHODS

This report examines a method for classifying building energy consumption and quantifying its characteristics using machine learning methods for a group of buildings with a mix of uses and multiple energy sources. As a result of this analysis, characteristic of building energy consumption could be distinguishable according to feature analysis based on the re-clustering method. Furthermore, the bubble chart diagram, which visually represents the building energy consumption efficiency and building energy consumption characteristics for each building, was shown to be effective for identifying priority countermeasure buildings in each cluster.