Spatial load forecasting based on generalized support vector machines and cellular automaton theory

A method of spatial load forecasting based on support vector machines (SVM) and cellular automata (CA) is proposed in this paper. Spatial load forecasting can be regarded as a complex problem considering many kinds of factors, while the theory of CA fully reflects the idea "complex systems are from the interaction of simple subsystem¿, and is an easy and effective method to deal with complicated problems. In the paper, in order to make the mode more suitable to simulate the space distribution of power load, the traditional mode of CA is advanced. Moreover, since the core definition of CA is to vary regulations, generalized SVM mode is applied to practice the conversion regulations of CA and it can effectively solve nonlinear problem and data noise question during the practice research of projects. The case in the paper proves the mode, put forward in the paper, accords with the factual situation of spatial load forecast in distribution power system.