The estimation of baseline electricity consumption for the implementation of energy efficiency and load management measures is an essential issue. In this study, two different forecasting methods for determining a consumption baseline are implemented and assessed: A first method based on Self-organizing Maps and a second one based on Principal Component Analysis. Both methods were applied to a set of data from an industrial site in France in order to evaluate their performance. For this evaluation, two specific criteria are proposed. Results show a good performance of both methods, but a better performance of the Self-organizing Maps based method. Perspectives for improving the performance of both methods through different adjustment factors are encouraging and contemplated for future works.