Power demand forecasting using stochastic model: Parameter estimation

Since it was first proposed, the new method of power demand forecasting using a stochastic model has been greatly talked about in the scientific community and increasing number of practical applications are being developed, since it is a simple and effective approach to forecast power demand within a small time interval. However, the relevant literature still assumes that all statistic parameters can be estimated perfectly and this assumption might not always be applicable. In this paper, we propose to use a shift register array to estimate these parameters. Using analysis techniques and factoring the length of the array we can then increase the precision in the estimated results. The precision of these results was proven using simulations, the results of which have also been documented here.