Data mining application: Classification of load pattern analysis for electricity customers

This paper deals with the wide range of models (Bayesian model, Support Vector Machines, Neural network) for forecasting the electricity. Electricity load forecasting has been object of vast research since energy load is known to be non-linear and, therefore, very difficult to predict with accuracy. Load forecasting is necessary for economic generation of power. Classification of load pattern is an important task for load forecasting of customers and grouping them into classes according to their load characteristics. The different clustering algorithms (modified follow-the-leader, k-means, fuzzy k-means and two types of hierarchical clustering) and the Self Organising Map to group together customers having a similar electrical behaviour. In the approach, all load curves of customers are first clustered with the clustering algorithms under a given number of clusters. This paper shows the study on Bayesian model, Support Vector Machines, Neural network and k-means algorithm.