Electricity Consumption Forecasting Using Facebook Prophet

Abstract

This research seeks for predicting electricity consumption in Baghdad Governorate using the Facebook Prophet Model, which was Proposed by Taylor and Latham in 2017 as a modern generalized additive Model specialized in tuning and forecasting time series data. This model includes non-linear functions that are combined into one model. The most important feature of this model is that it supports the presence of change points in the general trend, and uses the Fourier series to model seasonality, In addition, the model contains a function to represent the effect of special events (Holidays) in the time series. The electricity consumption data were analyzed and their estimated values were extracted according to the Facebook Prophet model using the Limited Memory Broyden, Fletcher, Goldfarb, and Shanno (L-BFGS) algorithm. After that, the estimated values were compared with the original values to measure the accuracy of the model`s performance. The results showed that the model provided relatively accurate prediction performance, which reflects the model`s ability to predict electricity consumption, and the possibility of using it as a tool to support energy demand management.