Energy Decomposition Model Using Takagi-Sugeno Neuro Fuzzy

Decomposition analysis is useful method to determine significant factors contribute towards the development of energy consumption. This paper presents factors decomposition of electricity consumption in Indonesia’s household sector using artificial intelligent method. The proposed artificial intelligent technique used in this study is the Neuro Fuzzy Takagi-Sugeno (NFTS) network, which is worked under multiple input multiple output condition. By tuning the appropriate Gaussian parameters, which are mean and variance, and two Takagi-Sugeno weight, the changes in electricity consumption that is decomposed into production effect, structural effect, and efficiency effect, has revealed. Compared to the common method, the performance of NFTS network for both constant and current price variables is quite satisfied, given the error generated in the network ranges between 0.003 and 2.09 %, which is quite low and acceptable.