Comparison of Forecasting Energy Consumption in Shandong, China Using the ARIMA Model, GM Model, and ARIMA-GM Model

Abstract

To scientifically predict the future energy demand of Shandong province, this study chose the past energy demand of Shandong province during 1995-2015 as the research object. Based on building model data sequences, the GM-ARIMA model, the GM (1,1) model, and the ARIMA model were used to predict the energy demand of Shandong province for the 2005-2015 data, the results of which were then compared to the actual result. By analyzing the relative average error, we found that the GM-ARIMA model had a higher accuracy for predicting the future energy demand data. The operation steps of the GM-ARIMA model were as follows: first, preprocessing the date and determining the dimensions of the GM (1,1) model. This was followed by the establishment of the metabolism GM (1,1) model and by calculation of the forecast data. Then, the ARIMA residual error was used to amend and test the model. Finally, the obtained prediction results and errors were analyzed. The prediction results show that the energy demand of Shandong province in 2016-2020 will grow at an average annual rate of 3.9%, and in 2020, the Shandong province energy demand will have increased to about 20% of that in 2015.

Incorporating causality in energy consumption forecasting using deep neural networks

“…These models are especially useful when there is heterogeneity in the time series (Bauwens et al, 2006 ; Engle, 2001 ). Examples of studies using ARMA/ARIMA models for predicting the future energy demands include Li and Li ( 2017 ) predicting the future energy consumption in the Shandong province in China; Ozturk and Ozturk ( 2018 ) predicting the coal, oil, natural gas, and renewable energy consumption in Turkey. Furthermore, Eerdogdu ( 2007 ) also used cointegratison analysis with ARIMA to predict total energy consumption in Turkey, while ) used ARIMA models to predict agricultural loads at small scales.…”

A Few-shot Learning Model based on a Triplet Network for the Prediction of Energy Coincident Peak Days

“…There is a tremendous amount of literature on forecasting electricity demands (Kandil, El-Debeiky, and Hasanien 2002;Li and Li 2017). Many different approaches have been applied to this problem from naive methods, e.g., autocorrelation with weather data (McSharry, Bouwman, and Bloemhof 2005), autoregressive integrated moving average (Li and Li 2017), to machine learning models, such as K-nearest neighbors (El-Attar, Goulermas, and Wu 2009), fuzzy models (Ying and Pan 2008), and artificial neural networks (Saini 2008). However, the prediction of the day of five coincident peaks (5CP) is not like the prediction of power demands (a regression problem).…”

Estimation and Forecasts for the Share of Renewable Energy Consumption in Final Energy Consumption by 2020 in the European Union

“…Energy consumption forecast studies were conducted using the ARIMA models, taking into account both autoregressive and mobile environments. For instance, this type of study has been used to forecast the annual energy consumption in Iran [36], to analyze energy consumption of buildings [37], to solve the long-term system load forecasting problem caused by load fluctuation in some developed cities [38], to forecast electricity demand [39], to compare the predictive accuracy of various uni-and multivariate models in forecasting international city tourism demand for Paris [40], to forecast China's primary energy consumption [41] and growth trends in renewable energy consumption of China [42], to forecast energy consumption and greenhouse gases emission of India [43], for forecasting national and regional energy demand [44], or to compare forecasting of energy consumption in Shandong, China [45]. Ozturk and Ozturk [46] used the autoregressive integrated moving average (ARIMA) model to forecast total of energy consumption and its structure in Turkey for the next 25 years.…”