Available online at www.HighTechJournal.org HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 306 ISSN: 2723-9535 The Use of Regression Method on Simple E for Estimating Electrical Energy Consumption Arnawan Hasibuan 1, 2* , Widyana Verawaty Siregar 3 , Muzamir Isa 1 , Eddy Warman 4, Roby Finata 5, M. Mursalin 6 1 School of Electrical System Engineering, University Malaysia Perlis, Malaysia. 2 Department of Electrical Engineering, Universitas Malikussaleh, Aceh Utara, Indonesia. 3 Department of Management, Universitas Malikussaleh, Aceh Utara, Indonesia. 4 Department of Electrical Engineering, Universitas Sumatera Utara, Indonesia. 5 Energy Transaction, ULP Sungai Penuh, PT. PLN (Persero), Indonesia. 6 Department of Mathematics Education, Universitas Malikussaleh, Aceh Utara, Indonesia. Received 05 April 2022; Revised 23 July 2022; Accepted 06 August 2022; Available online 20 August 2022 Abstract The continuous increase in population growth has an impact on the electrical energy supply. Based on this increase, electric power producers serve customers using proper forecasts. Therefore, it is a necessity to select the right calculation method with easy implementation. In this study, the population forecasts and economic growth calculations using the GT (Growth Trend) regression method development on Simple E were obtained for the year 2028. Furthermore, electricity consumption estimation was carried out using the DL (Double Log) regression method with growth trend, R, AR, DW, and t values of 6.63%, 0.993, 0.992, 1.21, and 2.18, respectively. The results show that estimated energy consumption was 6.63% annually, with the achievable amount for 2028 being 19,839.83 GWh. Keywords: Forecast of Electricity Demand; Energy; Population; Economy and Regression. 1. Introduction Currently, economic and socially driven community life activities are highly dependent on the availability of electricity supply [1, 2]. This dependence increases annually based on population growth, economic development, technological progress, and other social dynamics [3-5]. Therefore, it can be addressed through the provision of sufficient and reliable electricity supply at an affordable price [6]. A long-term electricity system development plan is required to achieve [7] the electricity demand for the next few years [8, 9] through calculations and forecasting [10-12]. During forecasting, it is better to use the routine method carried out by several electric power companies in the world [13]. This is expected to obtain an accurate, close to realization, and accountable output [14, 15]. Furthermore, outputs that deviate by being extremely high or low are detrimental to companies. Extremely high estimated output with low demand leads to overcapacity and overinvestment. Conversely, the reverse leads to a blackout due to insufficient power supply [16]. This study describes the development of an electrical energy forecasting method from Simple E application using regression calculations for North Sumatra Province, Indonesia, until 2028. * Corresponding author: arnawan@unimal.ac.id http://dx.doi.org/10.28991/HIJ-SP2022-03-06  This is an open access article under the CC-BY license (https://creativecommons.org/licenses/by/4.0/). © Authors retain all copyrights. https://creativecommons.org/licenses/by/4.0/ https://orcid.org/0000-0003-3864-9107 https://orcid.org/0000-0003-4610-7340 https://orcid.org/0000-0003-1083-065X https://orcid.org/0000-0003-3797-5656 HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 307 2. Materials and Methods Figure 1 shows the study location is North Sumatra, one of the 37 provinces in Indonesia. Figure 1. Map of North Sumatra, Indonesia The study flowchart is presented in Figure 2. Figure 2. Study flowchart HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 308 2.1. Forecasting Method The forecasting method is a technique for predicting future values based on mathematical and statistical data or information about the past and present [17-19]. In general, this method can be classified into two main groups namely:  Quantitative Method: The quantitative method obtains an estimate based on past quantitative data accompanied by a series of mathematical rules to predict future values for example the regression method [20].  Qualitative Method: The qualitative method obtains an estimate based on past qualitative data. Its forecast results depend on the compiler's intuitive thinking, opinions, knowledge and experiences. This method is usually used due to a lack of representative data for suitable mathematical models [21, 22]. 2.2. Regression Method Regression is a measure of the relationship between two or more variables expressed in terms of an equation or function [23, 24]. To determine this relationship (regression) a strict separation is required between symbol X and Y which represents independent and dependent variables, respectively [25]. Both variables are usually causal or have a causal relationship of mutual influence. Therefore, regression is a function L or the form of a particular function between dependent Y and independent X variables. Y = f(X) (1) According to Cénac et al. (2020) [26], regression methods can be divided into linear and nonlinear: Linear Regression In linear regression, the relationship between the independent (X) and dependent (Y) variables in a mathematical equation is a linear or straight line [27]. Linear regression consists of two types, namely:  Simple Linear Regression: The linear or simple linear regression is the simplest straight line or linear relationship between X and Y variables [28, 29]. Its mathematical equation is shown Equation 2: 𝑌 = 𝑓(𝑥) = 𝑎 + 𝑏𝑥 (2)  Multiple Linear Regression: For multiple linear regression, there is more than one independent variable (X) in a forecast. Its equation function is shown Equation 3: 𝑌 = 𝑓(𝑥) = 𝑎 + 𝑏1𝑥1 + 𝑏2𝑥2 + ⋯ + 𝑏𝑛𝑥𝑛 (3) Nonlinear Regression Non-linear regression is a relationship or function in which the independent X and dependent Y variables are factors with a certain rank and can either be denominators (fractional functions) or exponentials. 2.3. Statistical Indicators During the regression method, statistical indicators are used to describe the combined and individual level or degree of closeness between dependent Y and independent 𝑥1, 𝑥2, 𝑥3, … , 𝑥𝑛 variables. It can also apply to relationships between independent variables alone. Common uses of statistical indicators are described below:  Correlation Coefficient (R) Correlation is a measure of the relationship between two or more variables expressed using the correlation coefficient R as the degree of closeness or relationship level. During analysis, the dependence of one variable on another or vice versa is not important. Correlation methods are used for closeness measurement between independent X and dependent Y variables. According to Wagner (2019) [30], guidelines for providing interpretations of R are as follows: R = 0.00 - 0.199: very low; R = 0.20 - 0.399: low; R = 0.40 - 0.599: moderate; HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 309 R = 0.60 - 0.799: strong/close; R = 0.80 - 1,000: very strong.  R Square Determination Coefficient (R2) A squared correlation coefficient R obtains an R2 value called the coefficient of determination or determinant index. This states the relationship (percentage) between independent (X1, X2, X3, .... Xn) and dependent Y variables simultaneously. R2 values are between 0 and 1 (0 ≤ 𝑅2 ≤ 1).  The Coefficient of Determination Adjusted R-Square (AR) The adjusted R Square (AR) interpretation is the same as R2, however, its value increases or decreases with the addition of a new independent or dependent variable depending on correlation. A negative AR value is considered 0, which means that the independent variable has absolutely no relationship with the dependent variable. The Adjusted R Square (AR) closeness statistics are: R = 0.00 - 0.199: very low; R = 0.20 - 0.399: low; R = 0.40 - 0.5; 99: moderate; R = 0.60 - 0.799: strong/close; R = 0.80 - 1,000: very strong.  Value of t Statistics This is often referred to as the t-value for testing the degree of relationship closeness between dependent Y and independent X variables individually or partially (Y with X1), (Y with X2), (Y with X3), etc. The criteria for the degree of relationship closeness are as follows: | t | ≥ 2: Significant; 2> | t | ≥ 1: Admissible to use; | t | <1: Insignificant.  Durbin-Watson Statistics (DW) The Durbin-Watson indicator (DW) is used to determine a correlation between independent variables, namely between X1 and X2, X1 and X3, X2 and X3, etc. 1 ≤ DW ≤ 3 is the acceptable DW value, with the following criteria: DW = 2: No serial correlation; DW → 0: Positive correlation; DW → 3: Negative correlation. 2.4. Gross Domestic Product Gross Domestic Product (GDP) is the total goods and services production output expressed in units of currency from a region of the economy (country) without considering the production factor owner for a given period. The Gross Regional Domestic Product (GRDP) review is limited to a Province or Regency/City, at certain periods (for example 1 year or 1 quarter). 2.5. Simple E Application The Simple E application is based on statistical methods [31] that take advantage of existing functions [32] in Microsoft Excel. It was developed by Yamaguchi (2000) [33] from the Institute of Energy Economics (IEE) Japan. Furthermore, Simple E was used by the Directorate General of Electricity and the Ministry of Energy and Mineral Resources. It is an inserted Microsoft Excel module [34], placed in an Add-In consisting of three main parts namely:  Sheet Data The sheet data is used for input data namely, business (selling energy, contractual power and number of customers), past economic growth, population, and future forecast data. It has data coding or naming which starts at the time of input. HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 310  Sheet Model The sheet model contains statistical models (time series and regression) selected to calculate the estimated electricity demand. It also shows statistical coefficient indicators (R, R2, AR, DW, t-Value) due to program execution.  Sheet Simulation The sheet simulation contains a complete calculation result with the formed regression equation formulas and average growth rate. 3. Results 3.1. Forecast Stages Increased electricity consumption is generally influenced by several factors, namely population and economic growth. A forecast for these factors was first obtained before that of electricity consumption was carried out in North Sumatra until 2028. The step table for electricity consumption forecast based on influencing factors can be seen in Table 1. Table 1. Stages of forecasts Years Population growth Economic growth Electric Energy Consumption (GWH) 2004 12,165,423 83,328,948.58 4,450.76 2005 12,297,894 87,897,791.21 4,613.37 2006 12,431,808 93,347,404.39 4,717.81 2007 12,567,180 99,792,273.27 5,163.43 2008 12,704,025 106,172,360.10 5,757.84 2009 12,842,362 111,559,224.81 6,096.89 2010 12,982,204 118,718,902.74 6,636.45 2011 13,103,596 126,587,621.89 7,194.04 2012 13,215,401 134,461,505.43 7,809.32 2013 13,326,307 142,537,121.58 7,917.23 2014 13,556,968 ? 8,271.01 2015 13,937,797 ? 8,703.66 2016 14,102,911 ? 9,240.30 2017 14,262,147 ? 9,707.33 2018 14,415,391 ? 10,445.02 2019 14,562,549 ? ? 2020 ? ? ? 2021 ? ? ? 2022 ? ? ? 2023 ? ? ? 2024 ? ? ? 2025 ? ? ? 2026 ? ? ? 2027 ? ? ? 2028 ? ? ? Forecast regarding the factors include:  Forecast of population growth  Forecast of economic growth  Electricity consumption forecast Forecast of Population Growth The calculation using Simple E application can be seen in Table 2. HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 311 Table 2. Test results using Simple E Years Population Growth Regression SL TL TG 2004 12,165,423 12,165,423 12,165,423 2005 12,297,894 12,297,894 12,297,894 2006 12,431,808 12,431,808 12,431,808 2007 12,567,180 12,567,180 12,567,180 2008 12,704,025 12,704,025 12,704,025 2009 12,842,362 12,842,362 12,842,362 2010 12,982,204 12,982,204 12,982,204 2011 13,103,596 13,103,596 13,103,596 2012 13,215,401 13,215,401 13,215,401 2013 13,326,307 13,326,307 13,326,307 2014 13,556,968 13,556,968 13,556,968 2015 13,937,797 13,937,797 13,937,797 2016 14,102,911 14,102,911 14,102,911 2017 14,262,147 14,262,147 14,262,147 2018 14,415,391 14,415,391 14,415,391 2019 14,562,549 14,562,549 14,562,549 2020 14,725,480 14,725,480 14,741,406 2021 14,888,411 14,888,411 14,922,464 2022 15,051,343 15,051,343 15,105,750 2023 15,214,274 15,214,274 15,291,293 2024 15,377,205 15,377,205 15,479,119 2025 15,540,136 15,540,136 15,669,256 2026 15,703,067 15,703,067 15,861,734 2027 15,865,998 15,865,998 16,056,581 2028 16,028,930 16,028,930 16,253,826 From the calculation data listed in Table 2, it can be described in the form of a graph whose results can be shown in Figure 3. Figure 3. Graph of population test results The trend coefficient values obtained from the Simple E application are shown in Table 3. Table 3. Test coefficient values No Regression Method Trend Coefficient (%) Name Cede 1 Semi Log SL 1,21 / 1,07 2 Linear Trend TL 1,21 / 1,07 3 Growth Trend TG 1,21 / 1,23 0 2,000,000 4,000,000 6,000,000 8,000,000 10,000,000 12,000,000 14,000,000 16,000,000 18,000,000 P o p u la ti o n Years SL TL TG HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 312 From the trend coefficient values obtained, the best regression method used is Growth Trend [35]. This is due to the forecast percentage (1.23%) being close to the real data growth percentage (1.21%). Meanwhile, other methods only have a percentage growth value of 1.07%. Forecast of Economic Growth The calculation results using Simple E application can be seen in Table 4. Table 4. Test results using Simple E Years Economic Growth Regression (million Rupiah) SL TL TG 2004 83,328,948.6 83,328,948.6 83,328,948.6 2005 87,897,791.2 87,897,791.2 87,897,791.2 2006 93,347,404.4 93,347,404.4 93,347,404.4 2007 99,792,273.3 99,792,273.3 99,792,273.3 2008 106,172,360.1 106,172,360.1 106,172,360.1 2009 111,559,224.8 111,559,224.8 111,559,224.8 2010 118,718,902.7 118,718,902.7 118,718,902.7 2011 126,587,621.9 126,587,621.9 126,587,621.9 2012 134,461,505.4 134,461,505.4 134,461,505.4 2013 142,537,121.6 142,537,121.6 142,537,121.6 2014 149,126,136.1 149,126,136.1 151,352,206.6 2015 155,715,150.6 155,715,150.6 160,712,415.0 2016 162,304,165.2 162,304,165.2 170,651,457.0 2017 168,893,179.7 168,893,179.7 181,205,127.6 2018 175,482,194.2 175,482,194.2 192,411,435.4 2019 182,071,208.7 182,071,208.7 204,310,739.4 2020 188,660,223.3 188,660,223.3 216,945,894.1 2021 195,249,237.8 195,249,237.8 230,362,404.7 2022 201,838,252.3 201,838,252.3 244,608,589.9 2023 208,427,266.9 208,427,266.9 259,735,756.8 2024 215,016,281.4 215,016,281.4 275,798,385.0 2025 221,605,295.9 221,605,295.9 292,854,323.3 2026 228,194,310.4 228,194,310.4 310,964,997.9 2027 234,783,325.0 234,783,325.0 330,195,633.4 2028 241,372,339.5 241,372,339.5 350,615,488.1 From the calculation data listed in Table 4, it can be described in the form of a graph whose results can be shown in Figure 4. Figure 4. Graph of economic test results 0 50000000 100000000 150000000 200000000 250000000 300000000 350000000 400000000 P o p u la ti o n Years SL TL TG HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 313 The trend coefficient values obtained from Simple E application are shown in Table 5. Table 5. Test coefficient values No. Regression Method Trend Coefficient (%) Name Code 1 Semi Log SL 6,15 / 3,57 2 Linear Trend TL 6,15 / 3,57 3 Growth Trend TG 6,15 / 6,18 From the trend coefficient values obtained, Growth Trend is the best regression method used. This is due to an estimated percentage (6.18%) which approaches the real data percentage growth (6.15%). Meanwhile, other methods only have a percentage growth of 3.57%. Electricity Consumption Forecast After obtaining the population and economic growth forecast for up to 2028, electricity consumption was forecasted next as shown in the following Table 6. Table 6. Forecast of electricity consumption Years Population growth Economic growth Electric Energy Consumption (Gwh) 2004 12,165,423 83,328,948.58 4,450.76 2005 12,297,894 87,897,791.21 4,613.37 2006 12,431,808 93,347,404.39 4,717.81 2007 12,567,180 99,792,273.27 5,163.43 2008 12,704,025 106,172,360.10 5,757.84 2009 12,842,362 111,559,224.81 6,096.89 2010 12,982,204 118,718,902.74 6,636.45 2011 13,103,596 126,587,621.89 7,194.04 2012 13,215,401 134,461,505.43 7,809.32 2013 13,326,307 142,537,121.58 7,917.23 2014 13,556,968 151,352,206.62 8,271.01 2015 13,937,797 160,712,414.99 8,703.66 2016 14,102,911 170,651,457.00 9,240.30 2017 14,262,147 181,205,127.65 9,707.33 2018 14,415,391 192,411,435.44 10,445.02 2019 14,562,549 204,310,739.35 ? 2020 14,741,406 216,945,894.14 ? 2021 14,922,464 230,362,404.70 ? 2022 15,105,750 244,608,589.93 ? 2023 15,291,293 259,735,756.77 ? 2024 15,479,119 275,798,384.97 ? 2025 15,669,256 292,854,323.30 ? 2026 15,861,734 310,964,997.88 ? 2027 16,056,581 330,195,633.41 ? 2028 16,253,826 350,615,488.09 ? The calculation results using Simple E application can be seen in Table 7. HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 314 Table 7. Test results using Simple E Years Electric Energy Consumption (Gwh) Based on the Method Semi Log (SL) Linear Trend (TL) Growth Trend (GT) Double Log (DL) Least Square (LS) 2004 4,450.76 4,450.76 4,450.76 4,450.76 4,450.76 2005 4,613.37 4,613.37 4,613.37 4,613.37 4,613.37 2006 4,717.81 4,717.81 4,717.81 4,717.81 4,717.81 2007 5,163.43 5,163.43 5,163.43 5,163.43 5,163.43 2008 5,757.84 5,757.84 5,757.84 5,757.84 5,757.84 2009 6,096.89 6,096.89 6,096.89 6,096.89 6,096.89 2010 6,636.45 6,636.45 6,636.45 6,636.45 6,636.45 2011 7,194.04 7,194.04 7,194.04 7,194.04 7,194.04 2012 7,809.32 7,809.32 7,809.32 7,809.32 7,809.32 2013 7,917.23 7,917.23 7,917.23 7,917.23 7,917.23 2014 8,271.01 8,271.01 8,271.01 8,271.01 8,271.01 2015 8,703.66 8,703.66 8,703.66 8,703.66 8,703.66 2016 9,240.30 9,240.30 9,240.30 9,240.30 9,240.30 2017 9,707.33 9,707.33 9,707.33 9,707.33 9,707.33 2018 10,445.02 10,445.02 10,445.02 10,445.02 10,445.02 2019 12,577.98 10,879.48 11,139.16 11,257.94 11,396.83 2020 14,050.86 11,313.95 11,878.49 11,989.56 12,194.77 2021 15,820.59 11,748.41 12,665.94 12,768.71 13,051.97 2022 17,963.38 12,182.88 13,504.64 13,598.47 13,972.23 2023 20,579.28 12,617.34 14,397.93 14,482.14 14,959.56 2024 23,800.96 13,051.81 15,349.37 15,423.21 16,018.24 2025 27,806.19 13,486.27 16,362.74 16,425.41 17,152.81 2026 32,835.85 13,920.73 17,442.06 17,492.72 18,368.10 2027 39,220.11 14,355.20 18,591.64 18,629.36 19,669.21 2028 47,417.16 14,789.66 19,816.04 19,839.83 21,061.59 The statistical coefficient values obtained from Simple E application are shown in Table 8. Table 8. Test coefficient values No. Regression Method Regression Indicators Trend Coefficient Name Code R AR DW t-value 1 Semi Log $SL 0.959 0.952 0.29 2.18 6.28/16.33 2 Linier Trend $TL 0.989 0.988 0.65 2.18 6.28/3.54 3 Linier Growth $TG 0.989 0.988 0.65 2.18 6.28/6.61 4 Double Log $DL 0.993 0.992 1.21 2.18 6.28/6.63 5 Least Square $LS 0.989 0.988 0.65 2.18 6.28/7.27 From the statistical coefficient values, the Double Log is the best regression method used. This is shown in the coefficient values obtained, namely growth trend, R, AR, DW and t values of 6.63%, 0.993, 0.992, 1.21 and 2.18 respectively. 3.2. Comparison of Forecast Using Simple E Application and Forecast Obtained from PT PLN (Persero) A comparison of the PT PLN (Persero) forecasts and those obtained from Simple E application was carried out [36]. The data obtained is used for the realization of electricity consumption in North Sumatra region from 2016 to 2018 based on RUPTL PT PLN (Persero) during 2016-2025, and the Simple E application. Time series data is used to obtain forecasts, namely data on the realization of population, economic, and electricity consumption growth in the North Sumatra region from 2004 to 2015. A comparison of the forecast values from PT PLN (Persero) [37] and the Simple E application can be seen in Table 9. HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 315 The Double Log Regression (DL) method was used for Simple E application above and obtained R, AR, DW and t values of 0.986, 0.982, 1.03 and 2.31, respectively. From Table 9, the results are closer to the realization data with an average percentage value of 7.04%. Meanwhile, the PT PLN (Persero) forecast results had an average deviation of 13.31% from the realization data. Table 9. Comparison of forecast results Years Realization of Electric Energy consumption (Gwh) Comparison of Forecast Values (Gwh) Comparison of Forecast Percentages (%) RUPTL 2016-2025 Simple E RUPTL 2016-2025 Simple E 2016 9,240.30 9,918 8,527.42 7.33% 7.71% 2017 9,707.33 11,046 9,089.96 13.79% 6.36% 2018 10,445.02 12,410 9,709.98 18.81% 7.04% Average 13.31% 7.04% Table 10 describes the forecast for electricity consumption in North Sumatra, Indonesia until 2028. It can be seen that there is almost no difference between using the 2019-2028 RUPTL and Simple E. The only difference in the forecast starts from 2020 to 2028, which according to Simple E calculations is lower. Table 10. Electricity consumption forecast Years Electricity Consumption Forecast (GWh) RUPTL 2019-2028 Simple E 2004 4,450.76 4,450.76 2005 4,613.37 4,613.37 2006 4,717.81 4,717.81 2007 5,163.43 5,163.43 2008 5,757.84 5,757.84 2009 6,096.89 6,096.89 2010 6,636.45 6,636.45 2011 7,194.04 7,194.04 2012 7,809.32 7,809.32 2013 7,917.23 7,917.23 2014 8,271.01 8,271.01 2015 8,703.66 8,703.66 2016 9,240.30 9,240.30 2017 9,707.33 9,707.33 2018 10,445.02 10,445.02 2019 11,361.00 11,257.94 2020 12,210.00 11,989.56 2021 13,286.00 12,768.71 2022 14,656.00 13,598.47 2023 15,979.00 14,482.14 2024 17,007.00 15,423.21 2025 18,105.00 16,425.41 2026 19,381.00 17,492.72 2027 20,742.00 18,629.36 2028 22,194.00 19,839.83 4. Conclusion This study considers electricity demand until 2028 with a Simple E application using Regression in North Sumatra Province, Indonesia. There was a focus on three trends in forecasting, namely semi-logs, linear, and NAT trends. A model was calibrated with historical data from 2004 to 2018. Furthermore, during the regression process using double logs, the coefficient values obtained include growth trend, R, AR, DW, and t values of 6.63%, 0.993, 0.992, 1.21, and HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 316 2.18, respectively. Application using the Double Log Regression (DL) method obtained R, AR, DW, and t values of 0.986, 0.982, 1.03, and 2.31, respectively. This was closer to the realization data, with an average percentage value of 7.04%. Meanwhile, the forecast results carried out by PT PLN (Persero) had an average deviation of 13.31% from the realization data. Using a value of “t = 2”, it can be concluded that the relationship between electricity consumption, with population and economic growth was significant. The estimated energy consumption growth in North Sumatra was 6.63% annually, with a total of 19,839.83 GWh in 2028. Finally, further studies may be required to validate the model accuracy for different periods or extended sample sets. This is carried out by applying the model to different energy markets. In this study, the load demand accounted for expandable variables such as weather forecasting measures, renewable energy impact on power grids, production technology, new trends in distributed energy generation, and market incorporation. 5. Declarations 5.1. Author Contributions Conceptualization, A.H.; methodology, A.H., W.V.S., and E.W.; formal analysis, W.V.S., and M.; investigation, E.W.; data curation, R.F.; writing—original draft preparation, A.H., W.V.S., M.I., E.W., R.F., and M.; writing—review and editing, A.H., and M.I. All authors have read and agreed to the published version of the manuscript. 5.2. Data Availability Statement The data presented in this study are available in article. 5.3. Funding The authors received no financial support for the research, authorship, and/or publication of this article. 5.4. Acknowledgements The author expresses gratitude to fellow employees of PT. PLN, fellow lecturers from Malikusssaleh University, North Sumatra University, and North Sumatra Muhamadiyah University, for the support provided during the completion of this study. The gratitude is also conveyed to the Authority of PT. PLN North Sumatra Region Indonesia for Energy Regulations and the provision of data and advice. 5.5. Ethical Approval Not applicable. 5.6. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. 6. References [1] Yuan, J. H., Kang, J. G., Zhao, C. H., & Hu, Z. G. (2008). Energy consumption and economic growth: Evidence from China at both aggregated and disaggregated levels. Energy Economics, 30(6), 3077–3094. doi:10.1016/j.eneco.2008.03.007. [2] Koirala, B. P., Koliou, E., Friege, J., Hakvoort, R. A., & Herder, P. M. (2016). Energetic communities for community energy: A review of key issues and trends shaping integrated community energy systems. Renewable and Sustainable Energy Reviews, 56, 722–744. doi:10.1016/j.rser.2015.11.080. [3] Xie, F., Liu, Y., Guan, F., & Wang, N. (2020). How to coordinate the relationship between renewable energy consumption and green economic development: from the perspective of technological advancement. Environmental Sciences Europe, 32(1), 1–15. doi:10.1186/s12302-020-00350-5. [4] Chenghu, Z., Arif, M., Shehzad, K., Ahmad, M., & Oláh, J. (2021). Modeling the dynamic linkage between tourism development, technological innovation, urbanization and environmental quality: Provincial data analysis of China. International Journal of Environmental Research and Public Health, 18(16), 8456. doi:10.3390/ijerph18168456. [5] PéRez-Arriaga, I. J., Jenkins, J. D., & Batlle, C. (2017). A regulatory framework for an evolving electricity sector: Highlights of the MIT utility of the future study. Economics of Energy and Environmental Policy, 6(1), 71–92. doi:10.5547/2160-5890.6.1.iper. [6] Taneja, J. (2019). If You Build It, Will They Consume? Key Challenges for Universal, Reliable, and Low-Cost Electricity Delivery in Kenya. SSRN Electronic Journal, 491. doi:10.2139/ssrn.3310479. [7] Ouedraogo, N. S. (2017). Modeling sustainable long-term electricity supply-demand in Africa. Applied Energy, 190, 1047–1067. doi:10.1016/j.apenergy.2016.12.162. HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 317 [8] Raza, M. Y., & Lin, B. (2022). Analysis of Pakistan's electricity generation and CO2 emissions: Based on decomposition and decoupling approach. Journal of Cleaner Production, 359, 132074.doi:10.1016/j.jclepro.2022.132074. [9] Njeru, G., Gathiaka, J., & Kimuyu, P. (2020). Modelling and Forecasting Electricity Demand for Commercial and Industrial Consumers in Kenya to 2035. In European Scientific Journal ESJ, 16(13). doi:10.19044/esj.2020.v16n13p162. [10] Anand, A., & Suganthi, L. (2020). Forecasting of Electricity Demand by Hybrid ANN-PSO Models. Deep Learning and Neural Networks: Concepts, Methodologies, Tools, and Applications, 865-882. IGI Global, Pennsylvania, United States. doi:10.4018/978-1-7998-0414-7.ch048. [11] Angelopoulos, D., Siskos, Y., & Psarras, J. (2019). Disaggregating time series on multiple criteria for robust forecasting: The case of long-term electricity demand in Greece. European Journal of Operational Research, 275(1), 252–265. doi:10.1016/j.ejor.2018.11.003. [12] Williams, S., & Short, M. (2020). Electricity demand forecasting for decentralised energy management. Energy and Built Environment, 1(2), 178–186. doi:10.1016/j.enbenv.2020.01.001. [13] Jesmeen, M. Z. H., Hossen, J., & Aziz, A. B. A. (2021). Unsupervised Anomaly Detection for Energy Consumption in Time Series using Clustering Approach. Emerging Science Journal, 5(6), 840-854. doi:10.28991/esj-2021-01314. [14] Zhang, X. M., Grolinger, K., Capretz, M. A. M., & Seewald, L. (2018). Forecasting Residential Energy Consumption: Single Household Perspective. 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA). doi:10.1109/icmla.2018.00024. [15] Bouktif, S., Fiaz, A., Ouni, A., & Serhani, M. A. (2018). Optimal deep learning LSTM model for electric load forecasting using feature selection and genetic algorithm: Comparison with machine learning approaches. Energies, 11(7), 1636. doi:10.3390/en11071636. [16] Tetlock, P. E., & Gardner, D. (2016). Superforecasting: The art and science of prediction. Random House, New York City, United States. [17] Garland, T., & Ives, A. R. (2000). Using the past to predict the present: Confidence intervals for regression equations in phylogenetic comparative methods. American Naturalist, 155(3), 346–364. doi:10.1086/303327. [18] Briscoe, N. J., Elith, J., Salguero‐Gómez, R., Lahoz‐Monfort, J. J., Camac, J. S., Giljohann, K. M., … Guillera‐Arroita, G. (2019). Forecasting species range dynamics with process‐explicit models: matching methods to applications. Ecology Letters, 22(11), 1940–1956. doi:10.1111/ele.13348. [19] Petropoulos, F., Apiletti, D., Assimakopoulos, V., Babai, M. Z., Barrow, D. K., Ben Taieb, S., … Boylan, J. E. (2022). Forecasting: theory and practice. International Journal of Forecasting, 38(3), 705–871. doi:10.1016/j.ijforecast.2021.11.001. [20] Jenset, G. B., & McGillivray, B. (2017). Quantitative historical linguistics: A corpus framework. Oxford University Press, Oxford, United Kingdom. [21] Casarrubea, M., Aiello, S., Di Giovanni, G., Santangelo, A., Palacino, M., & Crescimanno, G. (2019). Combining Quantitative and Qualitative Data in the Study of Feeding Behavior in Male Wistar Rats. Frontiers in Psychology, 10. doi:10.3389/fpsyg.2019.00881. [22] Feng, L., Wang, Q., Wang, J., & Lin, K.-Y. (2022). A Review of Technological Forecasting from the Perspective of Complex Systems. Entropy, 24(6), 787. doi:10.3390/e24060787. [23] Norouzian, R., Plonsky, L. (2018). Correlation and Simple Linear Regression in Applied Linguistics. The Palgrave Handbook of Applied Linguistics Research Methodology. Palgrave Macmillan, London, United Kingdom. doi:10.1057/978-1-137-59900- 1_19. [24] Rahimi, R. (2017). Organizational Culture and Customer Relationship Management: A Simple Linear Regression Analysis. Journal of Hospitality Marketing and Management, 26(4), 443–449. doi:10.1080/19368623.2017.1254579. [25] Basha, L., & Hoxha, F. (2019). Kernel Estimation of the Baseline Function in the Cox Model. European Scientific Journal, 15(6), 105–118. doi:10.19044/esj.2019.v15n6p105. [26] Cénac, P., Godichon-Baggioni, A., & Portier, B. (2020). An efficient Averaged Stochastic Gauss-Newton algorithm for estimating parameters of nonlinear regressions models. arXiv preprint arXiv:2006.12920. doi:10.48550/arXiv.2006.12920. [27] Brown, S. H. (2009). Multiple linear regression analysis: a matrix approach with MATLAB. Alabama Journal of Mathematics, 34, 1-3. [28] Montgomery, D. C., Peck, E. A., & Vining, G. G. (2021). Introduction to linear regression analysis (6th Ed.). John Wiley & Sons, Hoboken, United States. [29] Kumari, K., & Yadav, S. (2018). Linear regression analysis study. Journal of the Practice of Cardiovascular Sciences, 4(1), 33. doi:10.4103/jpcs.jpcs_8_18. HighTech and Innovation Journal Vol. 3, No. 3, September, 2022 318 [30] Wagner, W., E. (2019). Using IBM® SPSS® statistics for research methods and social science statistics. Sage Publications, Newbury, United States. [31] Makridakis, S., Wheelwright, S. C., & Hyndman, R. J. (2008). Forecasting methods and applications. John Wiley & Sons, Hoboken, United States. [32] Deluca, D., Gasson, S., & Kock, N. (2006). Adaptations that Virtual Teams Make so that Complex Tasks Can Be Performed Using Simple E-Collaboration Technologies. International Journal of E-Collaboration (IJeC), 2(3), 65–91. doi:10.4018/jec.2006070104. [33] Yamaguchi, K. (2000). Operation Manual of Simple E. JICA Study Team, the Institute of Energy Economics, Tokyo, Japan. Available online: http://open_jicareport.jica.go.jp/pdf/11617180_04.pdf (accessed on May 2022). [34] Kriswanto, S. (2014). Electricity Demand Forecast Using the Simple E Application. Jakarta, Indonesia. (In Indonesian). [35] Senior, V. L., Evans, L. C., Leather, S. R., Oliver, T. H., & Evans, K. L. (2020). Phenological responses in a sycamore–aphid– parasitoid system and consequences for aphid population dynamics: a 20 year case study. Global change biology, 26(5), 2814- 2828. doi:10.1111/gcb.15015. [36] Smith, J., & Wallis, K. F. (2009). A simple explanation of the forecast combination puzzle. Oxford Bulletin of Economics and Statistics, 71(3), 331–355. doi:10.1111/j.1468-0084.2008.00541.x. [37] Pamuji, N. (2014). Business Plan for the Supply of Electricity (RUPTL 2014-2015) PT PLN (Persero), Jakarta, Indonesia. Available online: https://gatrik.esdm.go.id/assets/uploads/download_index/files/9e670-coffee-morning-ruptl-2015-2024.pdf (accessed on May 2022). https://doi.org/10.1111/gcb.15015