Prediction of electricity consumption for household categories R-1 based on average air temperature with a single input transfer function approach

The transfer function model is one of the quantitative prediction models that is often used for multivariate time series data prediction. This model combines several characteristics of regression analysis with ARIMA periodical characteristics. This research explained the application of a single input transfer function model in predicting the amount of electricity consumption (Yt) in Jombang Regency based on the influence of the average air temperature (Xt)). In conducting modelling, the data used electricity consumption and the average air temperature from January 2011 to June 2017. The stages that need to be done are to see the stationary data of the output series (Yt) and input series (Xt), determine the best ARIMA model output series (Yt) and input series (Xt), do prewhitening, determine the order (b, r, s) through cross correlations (CCF) plots, and determine the best model of the transfer function by describing the order (b, r, s) along with the sequence (nt). From the model produced it can be explained that the amount of electricity consumption in Jombang is influenced by the increase in electricity consumption 1 month before (Y_t-1^˙), minus 1202862.9 times the average air temperature this month (X_t^˙) minus the average air temperature 12 months ago o (X_t-12), minus the average air temperature 13 months ago (X_t-13^˙) plus the residual this month (a_t) minus 0.83127 times the residual 1 month ago (a_t-1).