UNIVERSITY OF SAN CARLOS

School of Engineering
Department of Mechanical and Manufacturing Engineering

PREDICTING ELECTRICITY DEMAND BASED ON SELECTED SOCIO-ECONOMIC

INDICATORS IN THE PHILIPPINES

In Partial Fulfillment of the Requirement for

ME 424M: Methods of Research
F 8:30-9:30 AM

Submitted to:
Dr. Lanndon Ocampo, PhD.

Submitted by:
De Guzman, Johannes Bruce L.
Roma, Allcris
Saut, Bonifacio R.

March 2017

1
 Table of Contents
ABSTRACT........................................................................................................................4
CHAPTER 1.......................................................................................................................5
THE PROBLEM AND ITS SCOPE.................................................................................5
INTRODUCTION.........................................................................................................5
RATIONALE.............................................................................................................5
THEORETICAL BACKGROUND.............................................................................8
Regression Analysis.............................................................................................8
Simple Linear Regression Model.........................................................................8
Multiple Linear Regression Model........................................................................9
Coefficient of Determination.................................................................................9
Interpreting p-Values..........................................................................................10
REVIEW OF RELATED LITERATURE..................................................................11
Forecasting Electricity Demand.........................................................................13
Regression Analysis...........................................................................................14
THE PROBLEM........................................................................................................16
STATEMENT OF THE PROBLEM.........................................................................16
SIGNIFICANCE OF THE STUDY..........................................................................17
RESEARCH METHODOLOGY.................................................................................18
Research Design...................................................................................................18
Research Environment..........................................................................................18
Research Participants............................................................................................19
Research Procedures............................................................................................19
DEFINITION OF TERMS..........................................................................................20
CHAPTER 2.....................................................................................................................22
PRESENTATION, INTERPETATION AND ANALYSIS OF DATA................................22
CHAPTER 3.....................................................................................................................30
SUMMARY OF FINDINGS, CONCLUSIONS, AND RECOMMENDATIONS..............30
SUMMARY OF FINDINGS........................................................................................30
CONCLUSIONS........................................................................................................30
RECOMMENDATIONS.............................................................................................31
REFERENCES................................................................................................................32
2
3
 ABSTRACT

In this study, the annual gross electricity demand of the Philippines was modeled

using multiple linear regression with the aid of Microsoft Excel as a function of three

chosen socio-economic factors namely, population, gross domestic product and inflation

rate. Among these three socio-economic factors, the inflation rate was found to be the

least statistically significant variable with a confidence interval of 72.8 percent while

population was found out to be the most statistically significant descriptor variable with a

confidence interval of 98.9 percent affecting the electricity demand for future

predictions. Next, the future values of the three chosen socio-economic factors were

predicted using the GMDH Shell DS software to forecast the future annual electricity

demand up to the year 2025. It was found that the electricity demand will increase up to

90,000 GWh from the year 2014 to 2025.

4
 CHAPTER 1

THE PROBLEM AND ITS SCOPE

INTRODUCTION

RATIONALE

Electricity surrounds us, running technology like the cellphones we are using,

computers, lights and any other devices that needs electricity in order for it to function.

Electricity plays a vital role to the advancement of technology and to the betterment of

peoples way of living around the world.

Electricity is most often generated at power station by electromechanical

generators, primarily driven by heat engines fuelled by combustion or nuclear fission but

also by other means such as kinetic energy of flowing water and wind. Other energy

sources include solar photovoltaic and geothermal power. The electricity that we use is

a secondary energy source because it is produced by converting primary sources of

energy such as coal, natural gas, nuclear energy solar energy and wind energy into

electrical power. It also referred to as energy carrier which means it can be converted to

other forms of energy such as mechanical energy or heat. Humans have an intimate

relationship with electricity to the point that it's virtually impossible to separate their lives

from it. As the world is evolving humans tend to nurture this intimate relationship thereby

increasing the amount of electricity demand for the following years to come.

This study can be a great help to the Philippines because the researchers want

to foresee on what will happen to the electricity demand in the Philippines for the

following years to come until 2025. The researchers also want to know what will be the

5
worst case scenario or best case scenario that will happen to the electricity

consumption on the said years to come. The aims of the researchers on forecasting the

electricity demand will be done by establishing a model using multiple linear regression

analysis with the consideration of the socio-economic factors that affect the electricity

demand namely the population, gross domestic product and inflation rate.

6
 THEORETICAL BACKGROUND

Regression Analysis

Regression analysis is a conceptually simple method for investigating functional

relationships among variables. The relationship is expressed in the form of an equation

or a model connecting the response or dependent variable and one or more explanatory

or predictor variables. The response variable is denoted by Y and the set of predictor

variables by X1 , X 2 , . . ., X p , where p denotes the number of predictor

variables. The true relationship between Y and X1 , X 2 , . . ., X p , can be

approximated by the regression model

(1)
Y =f ( X 1 , X 2 , , X p )+ ,

where E is assumed to be a random error representing the discrepancy in the

approximation. It accounts for the failure of the model to fit the data exactly. The function

f( X 1 , X 2 , . . ., X p ) describes the relationship between Y and X1 , X 2 , . . .,

Xp .

The predictor or explanatory variables are also called by other names such as
(2)
independent variables, covariates, regressors, factors, and carriers. The name

independent variable, though commonly used, is the least preferred, because in

practice the predictor variables are rarely independent of each other.

Simple Linear Regression Model

The simple linear regression model is a module with a simple regressor x that has a

relationship with a response y that is a straight line. The simple linear regression model

is given by (Montgomery et. al. 2013)

7
 Y = 0 + 1 X+ (2)

where the intercept 0 and the slope 1 are known constants with is a

random error component. The errors are assumed to have mean zero and unknown

2
variance . Additionally the errors are assumed to be uncorrelated. This means that

the value of one error does not depend on the value of any other error. The parameters

0 and 1 are usually called regression coefficients. These coefficients have a

simple and often useful interpretation. The slope 1 is the change in the mean of the

distribution of y produced by a unit change in x.

Multiple Linear Regression Model

A regression model that involves more than one regressor variable is called a multiple

regression model. The multiple linear regression model is (Montgomery et. al. 2013)

Y = 0 + 1 X 1+ 2 X 2 ++ p X p + , (3)

where y denotes the dependent variable, denotes the regression coefficients and

is the random error component. The term linear is used because equation 3 is a

linear function of the unknown parameters .

Coefficient of Determination

The coefficient of determination R2 is given by

SS R SS
R2= =1 E (4 )
SST SS T

where SS T is a measure of the variability in y without considering the effect of the

regressor variable x and SS T is a measure of the variability in y remaining after x has

8
been considered, R2 is often called the proportion of variation explained by the

regressor x. Values of R2 that are close to 1 imply that most of the variability in y is

explained by the regression model. The magnitude of R2 also depends on the range

of variability in the regressor variable. Generally R2 will increase as the spread of the

xs increases and decrease as the spread of the xs decreases provided the assumed

model form is correct.

Interpreting p-Values

Under the appropriate assumptions, the p-value is the conditional probability of

observing a value of the computed statistic. In some research areas, it has become

traditional to adopt a fixed significance level when examining p-values. The most

common choice for is 0.05. Acceptreject rules like this are generally unnecessary for

reasonable scientific inquiry. Simply reporting p-values and allowing readers to decide

on significance seems a better approach.

There is an important distinction between statistical significance, the observation

of a sufficiently small p-value, and scientific significance, observing an effect of sufficient

magnitude to be meaningful. Judgment of the latter usually will require examination of

more than just the p-value.

9
 REVIEW OF RELATED LITERATURE

According to Suhono and Sarjiya (2015) energy has significant role in the human

life. All of human actions need energy. Based on the fuel type, electricity has been the

most used after the oil. One of many reasons because of the electricity does not

produce emission when it used. Other advantages of electricity are easy to transmit,

easy to use and a lot of devices on earth need electricity as power source. Electricity

power, as a clean and efficient energy, is essential in our daily life. Compared with

traditional energy, electricity power is more suitable for the requirement of environment

friendly society. The electricity power system has played a key role in the economy

sector of a country (Wang et al. 2017). According to Vilar et al. (2009), nowadays, in

many countries all over the world, the production and sale of electricity is traded under

competitive rules in free markets. The agents involved in this market: system operators,

regulatory agencies, producers and consumers have a great interest in the study of

electricity load and price. Since electricity cannot be stored, the demand must be

satisfied instantaneously and producers need to anticipate to future demands to avoid

overproduction. Good forecasting of electricity demand is then very important for the

agents in the market.

According to Gnay (2016) the world electricity demand has increased extremely

in the recent years as the world has become more populated and as the electricity

consuming devices and appliances have become more common in the daily lives of

people. It is vital for a country to be able to supply the electricity exactly equal to the

demand. If the electricity generation capacity of a country is lower than the gross

demand, electricity dependent industry is affected negatively and blackouts occur; on

10
the other hand, a higher electricity generation capacity than the demand leads for the

power plants to work with idle capacity, which is a waste in economic resources. Hence,

accurate prediction of the electricity demand for the future is very important to correctly

plan and develop new electricity generation investments for maintaining the electricity

demandsupply balance. In order to forecast the electricity demand with a good

precision, one must correctly determine the variables which may influence the electricity

demand in that country. Population is one of the key factors that are highly correlated

with the electricity demand (more people consume more electricity). However,

population alone is not sufficient to explain the changes in the electricity demand

through years. It is also quite common to consider some economic indicators in

correlation with the electricity demand (Askarzadeh, 2014); one factor that can be used

for this purpose is the gross domestic product (GDP) per capita, which is an indicator of

the wealth of the people living in a country (Kucukali and Baris, 2010). As the GDP per

capita increases the living standards of people get better and their lifestyles become

more dependent on energy consuming devices and appliances. In addition to GDP per

capita, employment and the inflation rates are two other economic factors that may

affect the electricity demand (Zehadi et al. 2013). However, in this study only three

selected socio economic factors that affect electricity demand were considered namely

the population, gross domestic product and inflation rate. According to Chang et. al

(2016) several methods have been used to forecast electricity consumption over the last

few decades. These forecasting methods can be approximately classified into three

categories: causal models, time series analysis, and artificial intelligence approaches.

Causal models probe the relationships among multiple variables and assume that the

11
variations in dependent variables can be explained by independent variables;

specifically, historical data are used to establish a multivariate model for dependent

variable forecasting (Hair, 2010). The forecasting accuracy of a causal model depends

on the selection of independent variables. If the selected independent variables cannot

effectively explain the variation in the dependent variables, an inaccurate forecast is

produced.

According to Bisgaard et. al (2011) time series models include linear regression

and autoregressive integrated moving average analysis, time series models require only

historical observations to construct a model for forecasting the development of data

trends and are commonly used in forecasting energy demand. However, they require

high quantities of samples for accurate forecasting. For many years, many different

forecasting tools based on data mining were applied to predict the future electricity or

energy demand. Multiple linear regressions (MLR), methodologies based on fuzzy logic,

autoregressive forecasting methods, support vector regression models, and artificial

neural network (ANN) based models have been widely applied for these purposes. On

the other hand only one approach was considered in this study which is only the

multiple linear regression analysis.

Forecasting Electricity Demand

According to Boroojeni et al. (2017) electricity demand forecasting plays a pivotal

role in power systems management, especially for operation and maintenance

purposes. It is particularly more important for deregulated power systems, where the

forecast inaccuracies have significant implications for market operators, transmission

12
owners, and market participants. Forecasting demand is both a science and an art.

Econometric methods of forecasting, in the context of energy demand forecasting, can

be described as the science and art of specification, estimation, testing and evaluation

of models of economic processes that drive the demand for fuels. The need and

relevance of forecasting demand for an electric utility has become a much-discussed

issue in the recent past. This has led to the development of various new tools and

methods for forecasting in the last two decades.

According to de Castro and Cramton (2009), an essential step of resource

planning in electricity markets is assuring that there will be sufficient resources to meet

future demand. Building capacity is costly and takes time. However, the economic

consequences of an electricity shortage may be severe. Thus, when setting capacity

targets the regulator must balance the cost of excess capacity against the cost of

shortage. The question we address here is how we can get an accurate prediction of

the future demand.

Regression Analysis

According to Yukseltan et al. (2017) regression methods that are quite

convenient and easy to implement have been applied widely for electricity demand

forecasting. Since the early 1970s, several studies on energy demand have been

performed using various estimation methods. Many studies have aimed to evaluate the

impact of economic activity and energy planning on energy demand. In recent years,

because predictive models are of vital importance for policymakers, they have used

these models to forecast and model energy consumption and demand (Tutun et al.

13
2015). According to Fan et al. (2017) multiple studies have highlighted the key role of

modeling and hence forecasts of future electricity demand, and particularly peak

demand play in grid planning and generation investment. A subset of this work focuses

on modeling and forecasting residential peak demand, and a related area of forecasting

involves assessing the potential impact of different types of interventions on this peak.

However, much of this work has been forced to utilize less than ideal approaches due to

the limitations of available data, particularly temporal and geographical data resolution.

In turn, this has limited the work assessing possible peak demand interventions and

highlights the potential value of demand models based on fundamental household

interval metering and survey data to simulate different possible interventions. There are

varied objectives and hence approaches to demand forecasting. Short-term electricity

demand forecasting is the key to generation and network operation, whilst longer-term

fore-casts serve various planning roles. Geographically limited forecasts primarily serve

network decision making whilst system-wide fore-casts have both network and

generation adequacy roles. Finally, there are sectorial studies focused on peak demand

estimation broken down by residential, commercial and industrial end-users. In this

study long-term forecasting electricity demand will be the one to be focused on since it

is essential to know what will happen to the future electricity demand in order to be

prepared for what will be the best case scenario and worst case scenario that might

happen.

14
15
 THE PROBLEM
STATEMENT OF THE PROBLEM

This study aims to predict the electricity demand in the Philippines several years

from now.

This study specifically answers to the following questions:

1 In the over-all electricity demand, how much comes from the industrial and the

private sectors?

2 Up to what quantified extent will be the electricity demand in the Philippines

several years from now?

3 What possible actions can be done to address the increasing electricity demand

here in the Philippines?

4 With the forecasting models presented, what should be the course of policy

making with regards to the aggregate energy demand?

16
 SIGNIFICANCE OF THE STUDY

The results of this research study will greatly contribute to address the issue of

increasing electricity demand in the Philippines. Forecasting the electricity demand will

not just simply show results, but will also give us an implication on what course of action

should be taken up to meet the increasing demands. Given the limited studies with

regards to electricity demand forecasting, this study will give the possible best and worst

case scenarios particularly in the energy crisis that we might experience in the future.

As population density increases over a period of time, energy demand

consequently, also increases. Predicting the electricity demand would therefore be a

useful tool in devising specific strategies to counter the adverse effects increasing

human population. Moreover, having a data on electricity demand will be crucial

especially to sectors involved in the production of such, regardless of the method of

producing electricity.

The results of this study will also be an important asset to the Philippine

government particularly the department of energy as they should also know the

prospect of electricity demand here in the Philippines in order to ensure stability in the

industry since electricity is known vital key to industry operation. The number of

investors planning to establish an electricity-producing firm here in the country must

also be at par with electricity demand.

17
 RESEARCH METHODOLOGY

Research Design

The quantitative approach of research was applied in this study. In order to

predict the electricity demand with a good precision, one must correctly determine the

key variables which may give a significant impact on the electricity demand in the

country. Since this paper focuses on predicting the electricity demand in the Philippine

setting, it is necessary to acquire data in the Philippines itself. Population is known to be

one of the key factors that is highly correlated with electricity demand, by the simple fact

that the more people living in a country, the more the consumption of electricity.

However, population alone is inadequate to provide and explain the changes in the

electricity demand through the years. Due to this, other factors are to be considered. As

for this research paper, socio-economic key indicators are correlated with the electricity

demand. One factor that was used for the purpose of predicting electricity demand is

the Gross Domestic Product (GDP) per capita, which is an indicator of fortune of the

people living in a country. Another factor, the inflation rate, defined as the

percentage rate of change of a price index over time which might also influence the

electricity demand is also considered in this study.

Research Environment

This research focuses on predicting the electricity demand in the Philippines

years from now. Hence, having Philippines as the research locale.

18
Research Participants

The people within Philippine territory was the participants. Data of the past

electricity demand, population, GDP and the inflation were acquired from World Bank

Open Data, an online database giving free access to information such as those that

were aforementioned.

Research Procedures

After obtaining past data from the online database (World Bank Open Data),

GMDH Shell DS version 3.8.6, a software used for predicting the future values of the

selected socio-economic indicators; population, Gross Domestic Product (GDP) and

inflation rate was utilized. Afterwards, the data acquired was employed for the multiple

regression analysis, creating a correlation of the three socio-economic key factors to the

electricity demand. Microsoft Excel spreadsheet was used to analyze the data acquired

(Data Analysis tool). A model equation was attained after the multiple regression

analysis and by using this equation, a prediction of the electricity demand was

established until the year 2025.

19
 DEFINITION OF TERMS

1 Data Mining refers to the computational process of discovering patterns in large

data sets involving methods at the intersection of artificial intelligence, machine

learning, statistics, and database system.

2 Dependent Variable - represent the output or outcome whose variation is being

studied; a symbol that stands for an arbitrary output.

3 Electricity refers to the form of energy that drives the technology.
4 Electric Demand - refers to the maximum amount of electrical energy that is

being consumed at a given time. It is measured in both kilowatts and kilovolt

amperes, depending on the rate tariff.

5 Electromechanical Generators relates to mechanical human motion energy

converter to electrical energy that operates on the principle of electromagnetic

induction.
6 Forecasting process of making predictions of the future based on past and

present data and most commonly by analysis of trends.

7 Gross Domestic Product (GDP) - is the monetary value of all the finished goods

and services produced within a country's borders in a specific time period.

8 Heat Engines a system that converts thermal energy thermal energy and

chemical energy to mechanical energy, which can be used to do mechanical

work. It includes all private and public consumption, government outlays,

investments and exports minus imports that occur within a defined territory.
9 Independent Variable - is a variable that is being manipulated in an experiment in

order to observe the effect on a dependent variable.

10 Inflation - is the rate at which the general level of prices for goods and services is

rising and, consequently, the purchasing power of currency is falling.

11 Peak Demand refers to the period in which electrical power is expected to be

provided for a sustained period at a significantly higher than average supply

level.

20
12 Population - is the number of all the organisms of the same group or species,

which live in a particular geographical area.

13 Power Station refers to the place where electricity is being generated.
14 Prediction - is the process of determining the magnitude of statistical variates at

some future point of time, sometimes called an outcome variable.

15 R-squared Coefficient - is a statistical measure of how close the data are to the

fitted regression line. It is also known as the coefficient of determination, or the

coefficient of multiple determination for multiple regression.

16 Regression Analysis refers to the statistical process for estimating the

relationships among variables. It includes many techniques for modeling and

analyzing several variables, when the focus is on the relationship between a

dependent variable and one or more independent variables (or predictors).

17 Regression Coefficient - is the constant that represents the rate of change of one

variable as a function of changes in the other variable; it is the slope of

the regression line.

18 Socio-economic indicator are factors that affect the social and economic

development in a society.

21
 CHAPTER 2
PRESENTATION, INTERPETATION AND ANALYSIS OF DATA

Table 1. Electricity demand in the Philippines by sector (Philippine Power Statistics

2014).

Electricity Demand
Year Residenti Commerci Industri Other
Total
al al al s

200 3914
13,547 10,098 14,452 1,042
1 0

200 3862
13,715 10,109 13,628 1,172
2 4

200 4272
15357 11,106 15,188 1,069
3 0

200 4407
15920 11,785 15,012 1,359
4 6

200 4515
16,031 12,245 15,705 1,177
5 9

200 4567
15,830 12,679 15,888 1,275
6 2

200 4800
16,376 13,470 16,522 1,641
7 9

200 4920
16,644 14,136 17,031 1,395
8 6

200 5086
17,504 14,756 17,084 1,523
9 8

201 5526
18,833 16,261 18,576 1,596
0 6

201 5609
18,694 16,624 19,334 1,446
1 8

22
 201 5921
19,695 17,777 20,071 1,668
2 1

201 6156
20,614 18,304 20,677 1,971
3 6

201 6334
20,969 18,761 21,429 2,186
4 5

Figure 1. Plot of actual electricity demand in the Philippines.

Table 2. Selected socio-economic indicators in the Philippines (IEA Statistics 2014).

Description
Year Populati GDP per Inflation (%)
on Capita
796045
2001 41 958.01 5.35

812943
2002 78 1000.78 2.72

2003 829717 1011.29 2.29

23
 34

845962
2004 49 1080.09 4.83

861413
2005 73 1196.54 6.52

875928
2006 99 1395.21 5.49

889655
2007 08 1678.85 2.90

902971
2008 15 1929.13 8.26

916418
2009 81 1836.87 4.22

930389
2010 02 2145.24 3.79

945012
2011 33 2371.85 4.65

960173
2012 22 2604.66 3.17

975716
2013 76 2786.01 3.00

991386
2014 90 2873.09 4.10

24
Figure 2. Actual data of gross domestic product (GDP) per capita in the Philippines.

25
 Figure 3. Actual data of inflation rate in percentage.

Table 3. Regression statistics.

Variables Value
Multiple R -29778.87
2
R 0.983762564
Adjusted R2 0.978891333
Standard Error 1171.006542
Observations 14

The goodness of fit of a multiple linear regression model is measured by its

coefficient of determination. Referring to the data obtained shown in table 3, the

2
coefficient of determination R obtained after the regression analysis is 0.983762564.

26
The value of coefficient of determination which is close to 1 implies that the percentage

of variation of electricity demand is explained by the regression model. It should also be

noted that the value of coefficient of determination R2 depends on the range of

variability in the regressor variable.

Table 4. Parameters for estimating annual electricity demand.

Standard t p
Variables
Error Value Value
-
19465.75446
Intercept 29764.7548 -1.529 0.157
7
8
0.00081677
Population 0.000264 3.095 0.011
5
GDP per 4.27367114
2.344909 1.823 0.098
Capita 7
-
Inflation 233.900720 201.234507 -1.162 0.272
7

The data between the years 2001 and 2014 were used for multiple regression

analysis to construct an equation in order to relate the dependent variable to

independent (descriptor) variables, where electricity demand is the dependent variable,

while population , gross domestic product (GDP) per capita and inflation rate were the

independent variables. The regression coefficients are shown in table 4 which is used in

constructing the equation to be used in predicting the future values of electricity

demand. The equation obtained using the collected data through multiple regression is

given in equation 5.

y=29764.75488+0.000816775 x1 + 4.273671147 x 2233.9007207 x3 (5)

where:

27
 y = electricity demand
x 1 = population

x 2 = gross domestic product (GDP)

x 3 = Inflation rate

By examining the regression coefficients in table 2, the results show that population and

the gross domestic product (GDP) is directly proportional with electricity demand since

the coefficients are positive while the inflation rate is inversely proportional with the

electricity demand since it has a negative regression coefficient.

The statistical implication of each variable was determined by examining their

respective p values. It is evident by examining the p values of each variable that the

least statistically significant descriptor variable in performing future prediction of

electricity demand is the inflation rate with a confidence interval of 72.8 percent while

the most statistically significant variable is the population with a confidence interval of

98.9 percent.

Table 5. Predicted values of selected socio-economic indicators.

Description
Year Population GDP per Inflation Electricity
Capita (%) Demand
10072124
2015 6 3045.60 4.01 66441.78

10231657
68432.64
2016 2 3238.93 3.42

10392212
70799.79
2017 1 3472.09 3.67

10553261
73458.29
2018 4 3758.69 4.18

2019 10713985 4106.68 3.98 76211.47

28
 4

10873632
76335.03
2020 8 3824.24 4.09

11032100
78641.36
2021 7 4066.12 4.00

11190255
80980.13
2022 8 4339.88 3.47

11349605
83447.07
2023 7 4626.47 3.22

11511842
86547.04
2024 9 4861.97 6.51

11678465
90805.39
2025 3 5708.54 3.42

Figure 4. Predicted values of population for the years 2015-2025.

29
 Figure 5. Predicted values of inflation for the years 2015-2025.

Figure 6. Predicted values of gross domestic product (GDP) per capita.

30
 Figure 7. Predicted electricity demand based on selected descriptor variables.
The future values of population, GDP per capita and inflation rate were predicted

by GMDH Shell DS software. The plot of predicted values of population, GDP per capita

and inflation rate through years is shown in Figures 5, 6, 7 and 8. These figures

indicates that the population is expected to reach over 116 million in the year 2025,

while the GDP per capita is predicted to reach over $5700 in the same year. In addition

to these, the inflation is expected to decrease even though it was observed that it was

fluctuating on the previous predicted years. These values were used in order to

determine the trend of future electricity demand in the Philippines through multiple

regression analysis. The predicted values of selected socio-economic indicators and

electricity demand are shown in Table 5. It is evident in figure 8 that there will be a

continuous increase of electricity demand for the next years. In the year 2025 the

31
electricity demand as a function of the three selected socio-economic factors will reach

an estimated value of 90,000 GWh.

32
 CHAPTER 3

SUMMARY OF FINDINGS, CONCLUSIONS, AND RECOMMENDATIONS

SUMMARY OF FINDINGS

Table 1, shows the electricity demand in the Philippines, from each sectors dating

from 2001 to 2014 while Table 2 shows the values of the selected socio-economic

indicators in the Philippines for the same range of year (2001 to 2014). The data

gathered for the socio-economic indicators was used in obtaining the predicted values

of each until the year 2025. A correlation was then established using multiple regression

analysis to arrive at an equation that relates the dependent variables (socio-economic

indicators) to the independent variable which is the electricity demand.

From Figure 8, it can be clearly seen that trend of the graph continues to

increase, the population being the greatest contributor to the increase of electricity

demand, having a p value of 0.011, followed by the GDP with a p value of 0.098, then

followed by the inflation rate being the lowest factor to the increase in electricity demand

with a p value of (0.272).

CONCLUSIONS

In this paper, the electricity demand of Philippines was modeled using multiple

regression analysis using population, GDP per capita, and inflation rate as the

descriptor variables. Population and GDP per capita were found to be the major factors

that influence the demand while the inflation percentage indicates to be of minor

influence. With the increasing population every year, human needs particularly the

33
technological needs also increases. It would then consequently correspond to an

increase in electricity demand. In this study the electricity demand of the Philippines

from 2001 up to 2014 was modeled by multiple regression in Microsoft excel using the

socio-economic indicators that affect the electricity demand in the Philippines namely

the, population, gross domestic product per capita and inflation rate. The generated

equation for the multiple regression model was then used to predict the electricity

demand from 2015 up to 2025 and was found out that population being the key factor to

be considered in predicting electricity demand. Having known the predicted electricity

demand until year 2025, can be used to maintain the balance between electricity

demand and supply.

RECOMMENDATIONS

1. It is suggested to explore other socio-economic factors that might also

influence the increase of electricity demand for the years to come.

2. Philippines, being situated in the tropical zone, factors like average summer

temperature is recommended to be explored since it might be a significant

factor for the increase in electricity demand in the country.

34
 REFERENCES

Askarzadeh, A. (2014).Comparison of particle swarm optimization and other

metaheuristics one electricity demand estimation: a case study of Iran. Energy,

72, 484 491.

Bisgaard, S., and Kulahci, M. (2011). Time Series Analysis and Forecasting by

Example, John Wiley & Sons, Hoboken, NJ.

Boroojeni, K. G., Amini, H. M., Bahrami, S., Iyengar, S. S., Sarwat, A. I., and

Karabasoglu, O. (2017). A novel multi-time-scale modelling for electric power

demand forecasting: From short-term to medium-term horizon. Electric Power

Systems Research, 142, 58-73.

Chang, C., Lin, J., and Chang, M. (2016). Extended modeling procedure based on the

projected sample for forecasting short-term electricity consumption. Advanced

Engineering Informatics, 30, 211-217.

Chatterjee, S. and Hadi A. (2006). Regression Analysis by Example. John Wiley &

Sons, Inc., Hoboken, New Jersey.

De Castro, L., and Cramton, P. (2009). Prediction Markets to Forecast Electricity

Demand. 1-19.

Department of Energy (2014). Philippine Power Statistics. Retrieved from

https://www.doe.gov.ph/philippine-power-statistics.

Fan, H., MacGill, I.F., and Sproul, A. B. (2017). Statistical analysis of drivers of

residential peak electricity demand. Energy and Buildings, 141, 205-217.

35
Gnay, M. E. (2016). Forecasting annual gross electricity demand by artificial neural

networks using predicted values of socio-economic indicators and climatic

conditions: Case of Turkey. Energy Policy, 90, 92-101.

Hair, J.F. (2010). Multivariate Data Analysis: A Global Perspective. Pearson Education.

Kucukali, S., and Baris, K., (2010). Turkeys short-term gross annual electricity demand

forecast by fuzzy logic approach. Energy Policy, 38, 24382445.

Montgomery, D., Peck, E. and Vining, G. (2012). Introduction to Linear Regression

Analysis. John Wiley & Sons, Inc., Hoboken, New Jersey.

Suhono and Sarjiya (2014). Long-term electricity demand forecasting of Sumatera

System based on electricity consumption intensity and Indonesia population

projection 2010 2035. Energy Procedia, 69, 455-462.

Tutun, S., Chuo, C., and Canylmaz, E. (2015). A new forecasting framework for

volatile behaviour in net electricity consumption: A case study in Turkey. Energy,

93, 2405 2422.

Vilar, J.M., Cao, R., and Aneiros, G. (2009). Forecasting next-day electricity demand

and price using nonparametric functional methods. 1-23.

Wang, D., Luo, H., Grunder, O., Lin., Y., and Guo, H. (2017). Multi-step ahead

electricity price forecasting using a hybrid model based on two-layer

decomposition technique and BP neural network optimized by firefly algorithm.

Applied Energy, 190, 390 407.

Weisberg, S. (2005). Applied Linear Regression. John Wiley & Sons, Inc., Hoboken,

New Jersey.
36
Yukseltan, E., Yucekaya, A., and Bilge, A. H. (2017). Forecasting electricity demand for

Turkey: Modeling periodic variations and demand segregation. Applied Energy,

193, 287-296.

Zahedi, G.,Azizi, S.,Bahadori,A.,Elkamel,A.,WanAlwi,S.R. (2013). Electricity demand

estimation using an adaptive neuro-fuzzy network: a case study from the Ontario

province Canada. Energy 49,323328.

37