Analysis of the Behavior Pattern of Energy Consumption through Online Clustering Techniques

Analyzing energy consumption is currently of great interest to define efficient management strategies. In particular, studying the evolution behavior pattern can allow policies be defined according time year. this sense, work proposes study patterns using online clustering techniques. centroids groups constructed by techniques will represent their patterns. Specifically, two unsupervised machine learning ideal for stated objective analyzed, X-Means and LAMDA, since they are capable varying adapting number clusters at runtime. These applied data in commercial buildings, making groupings on previous groups, our case, monthly quarterly. We compared performance analyzing over time. The results very promising quality obtained good metrics. Thus, three main contributions article propose an approach determine non-supervised approaches, a methodology analyze explain clusters, comparison strategy have qualities order 0.59 0.41 Silhouette Davies-Boulding, respectively, 0.71 0.24 LAMDA different datasets energy. motivating terms particularly with LAMDA; therefore, its as user behaviors makes lot sense.