Long-term electric power load forecasting using fuzzy linear regression technique

This paper presents a new technique for long-term electric power load forecasting. The technique is based on fuzzy linear regression which uses long term annual growth factors to estimate fuzzy linear regression model parameters. In this technique a linear optimization problem is formulated, where the objective is to minimize the spread of fuzzy regression parameters. The annual growths for each of the long-term forecasting factors are calculated using cubic polynomials. The performance of the proposed technique is illustrated on real power network data.