FORECASTING MALAYSIA GROSS DOMESTIC PRODUCT BY USING LEAST SQUARE

METHOD
•
Nur Dini Huda binti Rosgi , Nur Hanani Sofiya Binti Mohamad Arif , and Faris Bin Nor Azmi
1
Faculty of Computer and Mathematical Sciences, Universiti Teknologi MARA (UiTM) Kampus Seremban
3,70300 Seremban, Negeri Sembilan

ABSTRACT – Malaysia's Gross Domestic Product (GDP) utilizing a static and

ARTICLE HISTORY
discrete mathematical model. The goal of this research is to figure out which
Received: 30-01-2020
model is best for forecasting GDP in the future. The issue we wish to investigate
Accepted: 05-03-2020
is which mathematical model, whether static or discrete, is the best for forecasting
GDP in the future.Hence, the problem we investigated is that GDP is a critical
KEYWORDS
component in determining a country's economic status. GDP is significant
Product Mix;
because it provides information on the size and performance of an economy. The
Linear Programming;
pace of increase in real GDP is frequently used as a gauge of the economy's
Small Medium
overall health. An increase in real GDP is viewed as an indication that the
Enterprise;
economy is performing well in general.Furthermore, we aim to investigate how
Drinking Water
important a mathematical model is and where errors in its use are most common.
This is what leads to inaccuracies in GDP forecasting in the future. To compare
errors, we need to find the RMSD for each model. A model for GDP forecast that
has a smaller RMSD value is more suited. As a result, we will compare the RMSD
value and choose the best representative model to offer as GDP prediction using
mathematical models such as static and discrete models.
INTRODUCTION

According to the International Monetary Fund, Malaysia's economy is the fourth largest in Southeast Asia
and the 38th largest in the world. Due to a high density of knowledge-based sectors and adoption of cutting-
edge technologies for manufacturing and the digital economy, labour productivity in Malaysia is much higher
than in neighbouring Thailand, Indonesia, the Philippines, or Vietnam. According to the Global
Competitiveness Report 2019, Malaysia's economy is the world's 27th most competitive, and it is one of the
top ten countries to embrace a digital legal framework.
Gross domestic product (GDP) is the total monetary or market value of all the finished goods and services
produced within a country’s borders in a specific time period. As a broad measure of overall domestic
production, it functions as a comprehensive scorecard of a given country’s economic health.
A mathematical model is a representation of a system that uses mathematical concepts and language to
describe it. It's a model that includes a set of system-related questions and a set of mathematical statements
that can be used to answer them.
Figure utilization, to provide uninterrupted production process for diverse customer demand in relation to
quality and timely delivery and to help the company to deliver a good products to customers on continuous
basis at competitive rates (Biswas & Chakraborty, 2016). One of the keys in production planning is the
determination of the optimal product mix as it can impact the company’s profitability. Product mix problem
involves the selection of optimal product combinations that maximize profit in demand constraints and
production resources (Gunasekaran, Zainali, & Aghapour, 2015). By achieving an optimal product mix, the
company is able to increase profits and minimize production costs. As stated by Ginting,

Figure 1: Classification of Mathematical Mode

Referring to above diagram, it shows that mathematical model is separated by two which are Static Model
(Least Square method) and Discrete Dynamical System. Static model is a model that does not provide time
dependency. We applied the Least Square method in the Static model. The Least Square model is a
mathematical regression analysis technique for determining the optimal line for a set of data. It may also use a
visual representation to highlight the link between data points. Each data point represents the relationship
between a known independent variable and an unknown dependent variable. There are three equations in the
Least Square method: a straight-line equation, a polynomial equation, and an exponential equation. The
straight-line equation is the common relationship between the x and y coordinates of any point on a line. Under
the static model, there are also polynomial equation and single exponential equation. A polynomial equation
is an equation with numerous terms made up of integers and variables. An exponential equation is a single-
exponent equation with a variable as the exponent.

Discrete model is a system which a quantity change value over discrete time interval. A nonlinear system
in which the country evolves in discrete time steps over state space following to a fixed rule is known as a
discrete dynamical system. One of the most prominent types of models in Discrete Dynamical Equations that
also we have been used in this project is the Discrete Malthusian Growth model. The Malthusian Growth
Model is named after Thomas Malthus, a British philosopher who lived in 1766. This model is also known as
the simple exponential growth model. This model is a model with only one variable and one parameter.

2
 The objective of this project are to determine the equation for each GDP percentage model. We want to
calculate the values of Root-Mean-Square Deviation (RMSD) for each model. Lastly, we are to determine the
most suitable model for the GDP percentage in Malaysia.

LITERATURE REVIEW

Gross Domestic Product

(Fernando, 2021) stated that the total monetary or market worth of all finished goods and services produced
inside a country's borders in a certain time period is known as GDP. It serves as a comprehensive scorecard of
a country's economic health because it is a wide measure of entire domestic production. GDP is normally
estimated on an annual basis, although it is also calculated on a quarterly basis. The government of the United
States, for example, publishes an annualized GDP estimate for each fiscal quarter as well as the calendar year.
Because the data in this report is presented in actual terms, it has been corrected for price fluctuations and is
thus inflation-adjusted.
(Collins & Nguyen, 2021) extend the micro to macro literature, according to D. Collins, by dissecting
earnings into R&D and pre-R&D components. They find that both components may forecast future real GDP
growth with varied lead-lag structures using Almon's finite distributed lag model (1965). Importantly, this
breakdown considerably improves the predictive model's explanatory ability when accounting data is used.
Through the GDP channels of personal consumption, business investment, and net export, aggregate
accounting R&D can predict real GDP. Their research builds on previous work on the forecasting utility of
accounting data at the aggregate level, and it has practical implications for macro forecasting and public policy
decisions involving publicly traded companies' creative activities. (Akgül et al., 2022) investigates four
fractional derivatives in the gross domestic product model. They also use the Sumudu transform to obtain
fractional model solutions and demonstrate the efficiency of the Sumudu transform through the use of theoretic
findings and applications.

3
 (Ge & Tang, 2020)explores the extent to which commodity prices can forecast GDP growth rates of various
countries using indices of 27 regularly traded commodity futures in another study linked to GDP prediction.
Commodity returns have a high prediction value for the next quarter's GDP growth, while the basis has a
moderate predictive capacity. Overall, commodity prices can be seen of as a leading indication of future
economic growth; rising commodity prices and basis values signal a stronger economy in the future. The futures
market can now give worldwide pricing information and has become an important indicator of economic
circumstances thanks to the creation of contemporary financial derivatives. Commodities play an essential role
in the economy as both raw materials for industrial production and important consumer goods for everyday
life. In the research of (Kouziokas, 2017), a machine learning methodology for time series forecasting of GDP
is proposed. To build forecasting models for projecting the Gross Domestic Product, Artificial Neural Networks
are used. Several network topologies were tested using different transfer functions and varied numbers
ofneurons in the hidden layers in order to produce the best forecasting model. The findings revealed that the
levels of Gross Domestic Product may be predicted with a high degree of precision.

(Zainon, 2019) built this study model using GDP data of Melaka and Negeri Sembilan for the years 2005-
2016. Discrete Dynamical System model is used to construct a mathematical model using state GDP data. The
states' future GDP was also predicted using the appropriate parameter value. In the long run, data that is not
monitored and examined will have an impact on economic growth. Any large shift in GDP, whether positive
or negative, can have a considerable impact on investor mood. If data is not monitored and examined, it is
obvious that economic activity and growth will suffer. To address this problem, a mathematical model will be
used to track and analyze data in an analytical and logical manner.

Static Model

Static and dynamic prediction models demonstrated excellent discriminations for incident Type 2 diabetes
(T2DM) in the (Asgari et al., 2021) study, with only minor variations between them. Our findings are
consistent with a study (Parastet al, 2019) that found that the dynamic land- mark model for predicting
incident T2DM had the same area under the curve (AUC) as the static model within three years among
people who were not taking metformin (dynamic = 0.682 vs. static = 0.678). (Russell et al., 2021) examines
the cost-effectiveness outcomes of two models of maternal immunization to prevent pertussis in infants in
Brazil, one static and the other dynamic, to see when static models are sufficient for public health choices and
when dynamic models are worth the extra work. The static model was also used to forecast the
costeffectiveness of maternal immunization under average conditions for the period 1999–2016, which is
similar to the dynamic model's projections(Russell et al., 2021).

4
Discrete Malthusian Model

According to previous studies, Malthusian forces played a significant effect in demographic,

socioeconomic, and political transformation. The Malthusian model is put to the test in this study by (Russell
et al., 2021), who uses data on storage capacity and population size from a single long-lived house in the Bridge
River town. The data in (Russell et al., 2021) research back up the claim that Malthusian processes have major
demographic, economic, and social consequences at the household level. (Jumarie, 2006) propose a new
solution for the stochastic differential equation of the Malthusian growth model with fractional random growth
rate, and apply it to the stability analysis of some nonlinear systems, following a review of fractional calculus
and a new derivation of Taylor's series of fractional order (the logistic law of growth, for instance).Most of the
researchers also have formulated a linear programming model with the objective function to maximize profit
under the capacity constraint, demand constraint and raw materials availability (Wang & Li, 2017). In product
mix problem, demand, cost and resource capacity are the most used constraints in the studies. In term of
resource capacity, raw material availability and machine capacity are usually being considered in previous
studies (Haider et al., 2016). From the previous research, it has helped the decision makers in the company in
making decision for their production planning where they focusing more on high-sales products for their
production (Maurya, Misra, Anderson, & Shukla, 2015).

The Company Case

The Gross Domestic Product of Malaysia was collected from the website Kaggle
(https://www.kaggle.com/holoong9291/gdp-of-all-countries19602020) based on our time series data. We
choose to use data from the Gross Domestic Product for a 20-year period, from 2001 to 2020.

Year GDP Percent (%)

2001 0.002803726
2002 0.002934542
2003 0.002856959
2004 0.002869697
2005 0.003047883
2006 0.003186091
2007 0.003362305
2008 0.003653648
2009 0.003375288
2010 0.003888252
2011 0.004089658
2012 0.004218684

5
 2013 0.004216296
2014 0.004290699
2015 0.004067341
2016 0.004001191
2017 0.003981883
2018 0.004211217
2019 0.004227818
2020 0.004341755

Figure 1: Years and Gross Domestic Product in Malaysia

RESULTS AND DISCUSSION

Static Model and Discrete Dynamical System are the two types of models that we used. We apply the
Least Square Method for the Static Model. The straight-line equation, polynomial equation, and single
exponential equation are the three equations we determine. We apply the Discrete Malthusian Growth Model
for the Discrete Dynamical System. The data for each model was likewise computed using Microsoft Excel.
In terms of data forecasting, we may utilize the equation of an appropriate model to meet the study's goals.
For each model, we 8 calculate the Root-Mean-Square Deviation (RMSD). Then, based on the lowest RMSD
value, we chose the optimal model.

Static Model

1. Straight Line Equation

The general form of Straight-line equation is, 𝑦 = 𝑎0 + 𝑎1 𝑥

Where,

y = independent variable (population)

𝑎0 ,𝑎1 = constant

𝑥 = dependent variable (year)

Year, x GDP (%) 𝑥2 𝑥𝑦

6
 1 0.002803726 1 0.0028

2 0.002934542 4 0.0059

3 0.002856959 9 0.0086

4 0.002869697 16 0.0115

5 0.003047883 25 0.0152

6 0.003186091 36 0.0191

7 0.003362305 49 0.0235

8 0.003653648 64 0.0292

9 0.003375288 81 0.0304

10 0.003888252 100 0.0389

11 0.004089658 121 0.0450

12 0.004218684 144 0.0506

13 0.004216296 169 0.0548

14 0.004290699 196 0.0601

15 0.004067341 225 0.0610

16 0.004001191 256 0.0640

17 0.003981883 289 0.0677

18 0.004211217 324 0.0758

19 0.004227818 361 0.0803

20 0.004341755 400 0.0868

∑𝑥 = 210 ∑𝑦 = 0.073624933 ∑𝑥 2 = 2870 ∑𝑥𝑦 = 0.8313

By using Camer’s rule, calculate 𝑎0 and 𝑎1 ,

𝑛 ∑𝑥 𝑎0 ∑𝑦
[ 2 ] [𝑎 ] = [ ]
∑𝑥 ∑𝑥 1 ∑𝑥𝑦

20 210 𝑎0 0.07
[ ][ ] = [ ]
210 2870 𝑎1 0.83

Value of |A| which is determinant of matrix,

20 210
𝐴=[ ]
210 2870

7
 Where |𝐴| = 13300

Obtain determinant 𝑎0 and 𝑎1 by using Cramer’s Rules,

|𝑎0 | = [0.07 210

]
0.83 2870

|𝑎0 | = 36.73

|𝑎1 | = [ 20 0.07
]
210 0.83

|𝑎1 | = 1.16

Calculate 𝑎0 and 𝑎1 using formula,

|𝑎0 | |𝑎1 |
𝑎0 = and 𝑎1 =
|𝐴| |𝐴|

Value for 𝑎0 and 𝑎1 which is,

|36.73|
𝑎0 =
|13300|

𝑎0 = 0.00276197

|1.16|
𝑎1 =
|13300|

𝑎1 = 0.00008755

Equation for straight line is,

𝑦 = 𝑎0 + 𝑎1 𝑥

𝑦 = 0.00276197 + 0.00008755𝑥

After that, we calculate the predicted Gross Domestic Product (GDP) by using the equation obtained above
and the square of Actual minus Predict as shown in the table below:

Actual Predict (Actual - Predict)2

0.002803726 0.00267441939 0.00000001672020
0.002934542 0.00258686920 0.00000012087638
0.002856959 0.00249931900 0.00000012790637
0.002869697 0.00241176880 0.00000020969823
0.003047883 0.00232421861 0.00000052369015
0.003186091 0.00223666841 0.00000090140325
0.003362305 0.00214911822 0.00000147182217

8
 0.003653648 0.00206156802 0.00000253471866
0.003375288 0.00197401782 0.00000196355811
0.003888252 0.00188646763 0.00000400714068
0.004089658 0.00179891743 0.00000524749236
0.004218684 0.00171136723 0.00000628663736
0.004216296 0.00162381704 0.00000672094717
0.004290699 0.00153626684 0.00000758689651
0.004067341 0.00144871665 0.00000685719351
0.004001191 0.00136116645 0.00000696972963
0.003981883 0.00127361625 0.00000733470877
0.004211217 0.00118606606 0.00000915153823
0.004227818 0.00109851586 0.00000979253188
0.004341755 0.00101096566 0.00001109415760
𝛴 (𝐴𝑐𝑡𝑢𝑎𝑙 − 𝑃𝑟𝑒𝑑𝑖𝑐𝑡)2
0.00008891936720

Calculate value RMSD by using formula,

Where n = 20 (number of years),

RMSD = 0.002108547

2. Single Exponential Equation

General Form of single exponential equation,

𝑦 = 𝐴𝑒𝐵𝑥

where,

= independent variable (population)

9
 = dependent variable (year)

= initial value of

= change in with respect to

Firstly, based on the equation, we need to linearize the equation as shown below,

𝑙𝑛 𝑦 = 𝑙𝑛 𝐴 + 𝑙𝑛 𝑒𝐵𝑥
𝑙𝑛 𝑦 = 𝑙𝑛 𝐴 + 𝑙𝑛 𝐵𝑥
Compare with
𝑦 = 𝛼0 + 𝛼
∴ 𝑌 = ln 𝑦
𝑎0 = ln 𝐴
𝑎1 = 𝐵
By using Least Square Method,

𝑦 = 𝛼0 + 𝛼1𝑥

Year, GDP (%) 𝑥2 𝑙𝑛 𝑦 𝑥(𝑙𝑛 𝑦)

1 0.002803726 1 -5.87681 -5.87681
2 0.002934542 4 -5.83120 -11.66241
3 0.002856959 9 -5.85800 -17.57399
4 0.002869697 16 -5.85355 -23.41420
5 0.003047883 25 -5.79331 -28.96654
6 0.003186091 36 -5.74896 -34.49376
7 0.003362305 49 -5.69513 -39.86590
8 0.003653648 64 -5.61203 -44.89623
9 0.003375288 81 -5.69127 -51.22147
10 0.003888252 100 -5.54980 -55.49796
11 0.004089658 121 -5.49929 -60.49223
12 0.004218684 144 -5.46823 -65.61878
13 0.004216296 169 -5.46880 -71.09438
14 0.004290699 196 -5.45131 -76.31828
15 0.004067341 225 -5.50477 -82.57149
16 0.004001191 256 -5.52116 -88.33861
17 0.003981883 289 -5.52600 -93.94201
18 0.004211217 324 -5.47000 -98.46006
19 0.004227818 361 -5.46607 -103.85532
10
 20 0.004341755 400 -5.43948 -108.78953
𝛴𝑦= 𝛴𝑙𝑛 𝑦 = - 𝑥(𝑙𝑛 𝑦) =-
𝛴𝑥 = 210 𝛴𝑥2 = 2870
0.073624933 112.32516 1162.94996

By using Cramer’s rule, calculate 𝛼0 and 𝛼1,

𝑛 ∑𝑥 𝑎0 ∑𝑙𝑛𝑦
[ ][ ] = [ ]
∑𝑥 ∑𝑥 2 𝑎1 ∑𝑥𝑙𝑛𝑦

20 210 𝑎0 −112.32516
[ ] [𝑎 ] = [ ]
210 2870 1 −1162.94996

Value of |Α| which is determinant of matrix,

where |Α| = 13300

Obtain determinant 𝛼0 and 𝛼1 by using Cramer’s Rules,

|𝛼0| = -78153.72

|𝑎1 | = [ 20 −112.32516
]
210 −1162.94996

|𝛼1| = 329.2847

Calculate 𝛼0 and 𝛼1 using formula,

and

Value for 𝛼0 and 𝛼1which is,

Equation exponential,

11
 𝑦 = 𝐴𝑒𝐵𝑥

𝐴0 = ln 𝐴

𝐴 = 0.0028054

𝐴1 = 𝐵

𝐵 = 0.0247583

𝑦 = 0.0028054𝑒 0.0247583𝑥

After that, we calculate the predicted Gross Domestic Product (GDP) by using the equation obtained above
and the square of Actual minus Predict as shown in the table below:

Actual Predict (Actual - Predict)2

0.002804 0.002875693 0.0000000051793021
0.002935 0.002947779 0.0000000001752231
0.002857 0.003021672 0.0000000271303711
0.00287 0.003097417 0.0000000518564437
0.003048 0.003175061 0.0000000161742251
0.003186 0.003254651 0.0000000047004837
0.003362 0.003336236 0.0000000006795756
0.003654 0.003419867 0.0000000546536936
0.003375 0.003505593 0.0000000169795150
0.003888 0.003593469 0.0000000868969184
0.00409 0.003683548 0.0000001649255988
0.004219 0.003775884 0.0000001960716636
0.004216 0.003870535 0.0000001195504254
0.004291 0.003967559 0.0000001044193608
0.004067 0.004067015 0.0000000000001062
0.004001 0.004168964 0.0000000281478094
0.003982 0.004273469 0.0000000850222127
0.004211 0.004380593 0.0000000286882059
0.004228 0.004490402 0.0000000689506134
0.004342 0.004602965 0.0000000682304931

𝛴 (𝐴𝑐𝑡𝑢𝑎𝑙 − 𝑃𝑟𝑒𝑑𝑖𝑐𝑡)2= 0.0000011284322408

12
 Calculate value RMSD by using formula,

Where n = 20 (number of years),

RMSD = 0.000237532

3. Polynomial Equation

Polynomial equation general form:

𝑦 = 𝑎0 + 𝑎1𝑥 + 𝑎2𝑥2

where,

𝑦 = independent variable (GDP percent)

𝑎0 , 𝑎1 , 𝑎2 = constant

𝑥 = dependent variable (year)

Year, 𝑥 GDP Percent (%) , 𝑦 𝑥2 𝑥3 𝑥4 𝒙𝒚 𝒙 𝟐𝒚

1 0.002803726 1 1 1 0.00 0.00

2 0.002934542 4 8 16 0.01 0.01

3 0.002856959 9 27 81 0.01 0.03

4 0.002869697 16 64 256 0.01 0.05

5 0.003047883 25 125 625 0.02 0.08

6 0.003186091 36 216 1296 0.02 0.11

7 0.003362305 49 343 2401 0.02 0.16

8 0.003653648 64 512 4096 0.03 0.23

9 0.003375288 81 729 6561 0.03 0.27

10 0.003888252 100 1000 10000 0.04 0.39

11 0.004089658 121 1331 14641 0.04 0.49

13
 12 0.004218684 144 1728 20736 0.05 0.61

13 0.004216296 169 2197 28561 0.05 0.71

14 0.004290699 196 2744 38416 0.06 0.84

15 0.004067341 225 3375 50625 0.06 0.92

16 0.004001191 256 4096 65536 0.06 1.02

17 0.003981883 289 4913 83521 0.07 1.15

18 0.004211217 324 5832 104976 0.08 1.36

19 0.004227818 361 6859 130321 0.08 1.53

20 0.004341755 400 8000 160000 0.09 1.74

𝛴𝑥 𝛴𝑦 𝛴𝑥2 𝛴𝑥3 𝛴𝑥4 𝛴𝑥𝑦 𝛴𝑥2y
=210 =0.073624933 =2870 =44100 =722666 =0.83 =11.71

By using Cramer’s Rules, calculate 𝑎0 , 𝑎1 and 𝑎2 from the data,

𝑛 ∑𝑥 ∑𝑥 2 𝑎0 ∑𝑦
[ ∑𝑥 ∑𝑥 2 ∑𝑥 ] [𝑎1 ] = [ ∑𝑥𝑦 ]
3

∑𝑥 2 ∑𝑥 3 ∑𝑥 4 𝑎2 ∑𝑥 2 𝑦

20 210.00 2870 𝑎0 0.07

[ 210 2870.00 44100 ] [𝑎1 ] = [ 0.83 ]
2870 44100.00 722666 𝑎2 11.71

Determinant, |A| calculated in matrix below,

| 𝐴 | = -54195320.00

Using Cramer’s Rules, determinant 𝑎0 , 𝑎1 and 𝑎2 were obtained,

0.07 210.00 2870

|𝑎0 | = [ 0.83 2870.00 44100 ]
11.71 44100.00 722666

|𝑎0| = - 566591.74.00

20 0.07 2870
|𝑎1 | = [ 210 0.83 44100 ]
2870 11.71 722666

|𝑎1| = - 8072721.41
14
 20 210.00 0.07
|𝑎2 | = [ 210 2870.00 0.83 ]
2870 44100.00 11.71

|𝑎2| = - 17270.96

𝑎0 , 𝑎1 and 𝑎2 were calculated using formula,

and

Equation for Polynomial;

𝑦 = 𝑎0 + 𝑎1 + 𝑎2𝑥2

𝑦 = −0.00104546 − 0.01489559𝑥 − 0.00003187𝑥2

Actual Predict (Actual-Predict)2

0.002803726 -0.02 0.000353

0.002934542 -0.03 0.001149

0.002856959 -0.05 0.002389

0.002869697 -0.06 0.004097

0.003047883 -0.08 0.006299

15
 0.003186091 -0.09 0.008978

0.003362305 -0.11 0.012153

0.003653648 -0.12 0.015852

0.003375288 -0.14 0.019899

0.003888252 -0.15 0.024673

0.004089658 -0.17 0.029875

0.004218684 -0.18 0.03557

0.004216296 -0.20 0.041734

0.004290699 -0.22 0.048453

0.004067341 -0.23 0.055562

0.004001191 -0.25 0.06327

0.003981883 -0.26 0.071536

0.004211217 -0.28 0.080487

0.004227818 -0.30 0.089876

0.004341755 -0.31 0.099885

∑(𝐴𝑐𝑡𝑢𝑎𝑙 − 𝑃𝑟𝑒𝑑𝑖𝑐𝑡)2 =0.71209

RMSD value were calculated using formula;

n = number of years

RMSD = 0.188691481

16
Discrete Dynamical System

1. Discrete Malthusian Growth Model

Assuming n is the GDP percentage and n is the year, the formula is as follows:

𝑃𝑛+1 = 𝑃𝑛 + 𝑟𝑃𝑛
= (1 + )𝑃𝑛
where;
= growth rate constant
= number of years
𝑃𝑛 = GDP percentage of
𝑃𝑛+1= Total GDP for year, +1

Year n GDP Percent r

2001 0 0.002803726 -

2002 1 0.002934542 0.04666

-
2003 2 0.002856959
0.02644

2004 3 0.002869697 0.00446

2005 4 0.003047883 0.06209

2006 5 0.003186091 0.04535

2007 6 0.003362305 0.05531

2008 7 0.003653648 0.08665

-
2009 8 0.003375288
0.07619

2010 9 0.003888252 0.15198

2011 10 0.004089658 0.05180

2012 11 0.004218684 0.03155

17
 -
2013 12 0.004216296
0.00057

2014 13 0.004290699 0.01765

-
2015 14 0.004067341
0.05206

-
2016 15 0.004001191
0.01626

-
2017 16 0.003981883
0.00483

2018 17 0.004211217 0.05759

2019 18 0.004227818 0.00394

2020 19 0.004341755 0.02695

We may get the growth rate constant, r, by substituting the data GDP percentage into this calculation. After
that, we calculate the average of r and construct the equation as follows:

Average of r

= 0.024507

The new equation is defined as:

𝑃𝑛 = (1 + ) 0

𝑃𝑛 = (1 + 0.1152 ) 0

𝑃𝑛 = 1.1152 0

Then, using the formula above, we calculate the predicted GDP percent and determine the percentage
inaccuracy after 2002. Fill in the value of n in the form for each year. Using the following formula, we can
calculate the percentage error for each year:

Percentage Error = x 100

18
 The table of calculation as shown below:

N Year Actual Predict %Re

1 2002 0.002934542 0.00301 -2.4507%
2 2003 0.002856959 0.00308 -7.8118%
3 2004 0.002869697 0.00316 -9.9636%
4 2005 0.003047883 0.00323 -6.0722%
5 2006 0.003186091 0.00331 -3.9577%
6 2007 0.003362305 0.00339 -0.9236%
7 2008 0.003653648 0.00348 4.8480%
8 2009 0.003375288 0.00356 -5.5234%
9 2010 0.003888252 0.00365 6.1531%
10 2011 0.004089658 0.00374 8.5882%
11 2012 0.004218684 0.00383 9.2123%
12 2013 0.004216296 0.00392 6.9346%
13 2014 0.004290699 0.00402 6.3072%
14 2015 0.004067341 0.00412 -1.2601%
15 2016 0.004001191 0.00422 -5.4568%
16 2017 0.003981883 0.00432 -8.5651%
17 2018 0.004211217 0.00443 -5.1686%
18 2019 0.004227818 0.00454 -7.3229%
19 2020 0.004341755 0.00465 -7.0676%

Then, as indicated below, we must calculate the value of Actual minus Predict:

N Year Actual Predict %Re Actual-Predict

1 2002 0.002935 0.00301 -2.4507% 2001.99707
2 2003 0.002857 0.00308 -7.8118% 2002.99714
3 2004 0.00287 0.00316 -9.9636% 2003.99713
4 2005 0.003048 0.00323 -6.0722% 2004.99695
5 2006 0.003186 0.00331 -3.9577% 2005.99681
6 2007 0.003362 0.00339 -0.9236% 2006.99664
7 2008 0.003654 0.00348 4.8480% 2007.99635
8 2009 0.003375 0.00356 -5.5234% 2008.99662

19
 9 2010 0.003888 0.00365 6.1531% 2009.99611
10 2011 0.00409 0.00374 8.5882% 2010.99591
11 2012 0.004219 0.00383 9.2123% 2011.99578
12 2013 0.004216 0.00392 6.9346% 2012.99578
13 2014 0.004291 0.00402 6.3072% 2013.99571
14 2015 0.004067 0.00412 -1.2601% 2014.99593
15 2016 0.004001 0.00422 -5.4568% 2015.99600
16 2017 0.003982 0.00432 -8.5651% 2016.99602
17 2018 0.004211 0.00443 -5.1686% 2017.99579
18 2019 0.004228 0.00454 -7.3229% 2018.99577
19 2020 0.004342 0.00465 -7.0676% 2019.99566

Where P0 = 0.002934542

Then, as stated in the table below, we must square the amount of Actual minus Predict and total it:

Actual-Predict (A-P)square
-0.00007 0.000000005171999819
-0.00022 0.000000049808727581
-0.00029 0.000000081753340308
-0.00019 0.000000034252396504
-0.00013 0.000000015900171796
-0.00003 0.000000000964314733
0.00018 0.000000031374605893
-0.00019 0.000000034756156759
0.00024 0.000000057239218681
0.00035 0.000000123360605024
0.00039 0.000000151037802064
0.00029 0.000000085489044958
0.00027 0.000000073237936550
-0.00005 0.000000002626827734
-0.00022 0.000000047670903558
-0.00034 0.000000116316703167
-0.00022 0.000000047375972756

20
 -0.00031 0.000000095850556225
-0.00031 0.000000094162160166
SUM 0.000001148349444276

Finally, we use the formula below to get the RMSD for the Discrete Malthusian Growth Model:

RMSD = 0.000245844
Results and Comparison

Model Static Model Discrete Dynamical

System

Method Straight Line Polynomial Single Discrete Malthusian

Equation Equation Exponential Growth

Equation

RMSD 0.002108547 0.188691481 0.000237532 0.000245844

As a consequence of the analysis, we can deduce that the Static Model by Single Exponential Equation is
the best model for forecasting Gross Domestic Product. Because the RMSD in single exponential equations is
smaller than the others which is 0.000237532.

21
CONCLUSION

In conclusion, mathematical models can assist us in comprehending and exploring the application of
equations or functional connections in this subject. Then we can learn how to use mathematical modeling tools
like Microsoft Excel. As a consequence, we can put what we've learned into practice in our everyday lives.
We may deduce from the models that the Static Model in polynomial equations is the best model since it has
the minimum RMSD. Using equations that we uncover, we can also anticipate the value of the GDP in
Malaysia.

REFERENCES

Gross Domestic Product (GDP) Definition. (2022). Investopedia.

https://www.investopedia.com/terms/g/gdp.asp#toc-what-is-gross-domestic-productgdp

Collins, D. W., & Nguyen, N. Q. (2021). Aggregate accounting research and development expenditures and
the prediction of real gross domestic product. Journal of Accounting and Public Policy, 106901.

https://doi.org/10.1016/j.jaccpubpol.2021.106901

Akgül, E. K., Jamshed, W., Nisar, K. S., Elagan, S., & Alshehri, N. A. (2022). On solutions of gross
domestic product model with different kernels. Alexandria Engineering Journal, 61(2), 1289–1295.
https://doi.org/10.1016/j.aej.2021.06.067

Fatinah, Z., Noraziah, M., Faezzah, M. D., & Rahela, A. R. (2019). INTRODUCTION. PROJECTION
ANALYSIS OF STATES GROSS DOMESTIC PRODUCT USING DISCRETE DYNAMICAL
SYSTEMS, 4(20), 1–2.

N. Kouziokas, G. (2017). Machine Learning Technique in Time Series Prediction of Gross Domestic
Product. ACM International Conference Proceeding Series, Part F1325.

Ge, Y., & Tang, K. (2020). Commodity prices and GDP growth. International Review of

Financial Analysis, 71, 101512. https://doi.org/10.1016/j.irfa.2020.101512

Ortega-Bastida, J., Gallego, A. J., Rico-Juan, J. R., & Albarrán, P. (2021). A multimodal approach for
regional GDP prediction using social media activity and historical information. Applied Soft
Computing, 111, 107693.

https://doi.org/10.1016/j.asoc.2021.107693

Russell, L. B., Kim, S. Y., Toscano, C., Cosgriff, B., Minamisava, R., Lucia Andrade, A., Sanderson, C., &
Sinha, A. (2021). Comparison of static and dynamic models of maternal immunization to prevent
infant pertussis in Brazil. Vaccine, 39(1), 158–166.

22
 https://doi.org/10.1016/j.vaccine.2020.09.006

Prentiss, A. M., Foor, T. A., & Hampton, A. (2018). Testing the Malthusian model: Population and storage at
Housepit 54, Bridge River, British Columbia. Journal of Archaeological Science:

Reports, 18, 535–550. https://doi.org/10.1016/j.jasrep.2018.02.015

23
APPENDIX

Figure 2: Table of Static Model

24
 Figure 3: Static Model Calculation

Figure 4: Discrete RMSD Calculation

ACKNOWLEDGEMENT

The authors would like to thank everyone that direct or indirectly gave a help for us in completing our mini
project.
25