This document analyzes the gross domestic product (GDP) of the United States, United Kingdom, and Australia using time series analysis. It finds that while the GDP of all three countries increases exponentially over time, there are also deviations from the trend. The document fits multivariate and univariate autoregressive models to the differenced GDP data to remove nonstationarity. It uses the models to make predictions for GDP over the next five years and compares the results across three statistical programs: ITSM, R, and SAS.
Splay Method of Model Acquisition Assessmentijtsrd
We know that the study of an object under investigation using a mathematical model is called modeling. The purpose of modeling will be to determine the properties of the object under study, to model it and to assess its condition. Turapov U. U | Isroilov U. B | Raxmatov A. SH | Egamov S. M | Isabekov B. I "Splay-Method of Model Acquisition Assessment" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-1 , December 2020, URL: https://www.ijtsrd.com/papers/ijtsrd38141.pdf Paper URL : https://www.ijtsrd.com/other-scientific-research-area/other/38141/splaymethod-of-model-acquisition-assessment/turapov-u-u
Ricci Solitons in an Kenmotsu Manifold admitting Conharmonic Curvature Tensorrahulmonikasharma
The object of the present paper is to study Ricci solitons in an (?)-Kenmotsu manifold. In this paper, some curvature conditions of conharmonic curvature tensor and pseudo-projective curvaturetensor have been studied. Under these conditions taking ? as space-like or time-like vector field, it is shown that Ricci solitons are expanding, steady or shrinking according as ? is positive, zero or negative respectively.
Splay Method of Model Acquisition Assessmentijtsrd
We know that the study of an object under investigation using a mathematical model is called modeling. The purpose of modeling will be to determine the properties of the object under study, to model it and to assess its condition. Turapov U. U | Isroilov U. B | Raxmatov A. SH | Egamov S. M | Isabekov B. I "Splay-Method of Model Acquisition Assessment" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-1 , December 2020, URL: https://www.ijtsrd.com/papers/ijtsrd38141.pdf Paper URL : https://www.ijtsrd.com/other-scientific-research-area/other/38141/splaymethod-of-model-acquisition-assessment/turapov-u-u
Ricci Solitons in an Kenmotsu Manifold admitting Conharmonic Curvature Tensorrahulmonikasharma
The object of the present paper is to study Ricci solitons in an (?)-Kenmotsu manifold. In this paper, some curvature conditions of conharmonic curvature tensor and pseudo-projective curvaturetensor have been studied. Under these conditions taking ? as space-like or time-like vector field, it is shown that Ricci solitons are expanding, steady or shrinking according as ? is positive, zero or negative respectively.
MIXTURES OF TRAINED REGRESSION CURVES MODELS FOR HANDWRITTEN ARABIC CHARACTER...gerogepatton
In this paper, we demonstrate how regression curves can be used to recognize 2D non-rigid handwritten shapes. Each shape is represented by a set of non-overlapping uniformly distributed landmarks. The underlying models utilize 2nd order of polynomials to model shapes within a training set. To estimate the regression models, we need to extract the required coefficients which describe the variations for a set of shape class. Hence, a least square method is used to estimate such modes. We proceed then, by training these coefficients using the apparatus Expectation Maximization algorithm. Recognition is carried out by finding the least error landmarks displacement with respect to the model curves. Handwritten isolated Arabic characters are used to evaluate our approach.
Bayesian Estimation for Missing Values in Latin Square Designinventionjournals
International Journal of Mathematics and Statistics Invention (IJMSI) is an international journal intended for professionals and researchers in all fields of computer science and electronics. IJMSI publishes research articles and reviews within the whole field Mathematics and Statistics, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online.
Numerical simulation of marangoni driven boundary layer flow over a flat plat...eSAT Journals
Abstract
A numerical algorithm is presented for studying Marangoni convection flow over a flat plate with an imposed temperature
distribution. Plate temperature varies with x in the following prescribed manner: where A and k are constants.
By means of similarity transformation, the original nonlinear partial differential equations of flow are transformed to a pair of
nonlinear ordinary differential equations. Subsequently they are reduced to a first order system and integrated using Newton
Raphson and adaptive Runge-Kutta methods. The computer codes are developed for this numerical analysis in Matlab
environment. Velocity profiles for various values of k, and temperature profiles for various Prandtl number and k are illustrated
graphically. The results of the present simulation are then compared with the previous works available in literature with good
agreement.
Keywords: Matlab, Marangoni Convection, Numerical Simulation, Surface Tension, Flat Plate.
This presentation covered the following topics :
1. Variance
2. Standard Deviation
3. Meaning and Types of Skewness
4. Related Examples
and is useful for B.Sc & M.Sc students.
Solving Transportation Problems with Hexagonal Fuzzy Numbers Using Best Candi...IJERA Editor
In this paper, we introduce a Fuzzy Transportation Problem (FTP) in which the values of transportation costs are
represented as hexagonal fuzzy numbers. We use the Best candidate method to solve the FTP. The Centroid
ranking technique is used to obtain the optimal solution.
Economics
Curve Fitting
macroeconomics
Curve fitting helps in capturing the trend in the data by assigning a single function
across the entire range.
If the functional relationship between the two quantities being graphed is known to be
within additive or multiplicative constants, it is common practice to transform the data in
such a way that the resulting line is a straight line.(by plotting) A process of quantitatively
estimating the trend of the outcomes, also known as regression or curve fitting, therefore
becomes necessary.
For a series of data, curve fitting is used to find the best fit curve. The produced equation is
used to find points anywhere along the curve. It also uses interpolation (exact fit to the data)
and smoothing.
Some people also refer it as regression analysis instead of curve fitting. The curve fitting
process fits equations of approximating curves to the raw field data. Nevertheless, for a
given set of data, the fitting curves of a given type are generally NOT unique.
Smoothing of the curve eliminates components like seasonal, cyclical and random
variations. Thus, a curve with a minimal deviation from all data points is desired. This
best-fitting curve can be obtained by the method of least squares.
What is curve fitting Curve fitting?
Curve fitting is the process of constructing a curve, or mathematical functions, which possess closest
proximity to the series of data. By the curve fitting we can mathematically construct the functional
relationship between the observed fact and parameter values, etc. It is highly effective in mathematical
modelling some natural processes.
What is a fitting model?
A fit model (sometimes fitting model) is a person who is used by a fashion designer or
clothing manufacturer to check the fit, drape and visual appearance of a design on a
'real' human being, effectively acting as a live mannequin.
What is a model fit statistics?
The goodness of fit of a statistical model describes how well it fits a set of
observations. Measures of goodness of fit typically summarize the discrepancy
between observed values and the values expected under the model in question.
What is a commercial model?
Commercial modeling is a more generalized type of modeling. There are high
fashion models, and then there are commercial models. ... They can model for
television, commercials, websites, magazines, newspapers, billboards and any other
type of advertisement. Most people who tell you they are models are “commercial”
models.
What is the exponential growth curve?
Growth of a system in which the amount being added to the system is proportional to the
amount already present: the bigger the system is, the greater the increase. ( See geometric
progression.) Note : In everyday speech, exponential growth means runaway expansion, such
as in population growth.
Why is population exponential?
Exponential population growth: When resources are unlimited, populations
exhibit exponential growth, resulting in a J-shaped curve.
Rajshahi Krishi Unnayan Bank is playing a vital role in the economic development of Bangladesh, especially in supporting farmers in sixteen districts of Rajshahi and Rangpur divisions. Agriculture is the foremost important part of the Bangladeshi economy.
Please Subscribe to this Channel for more solutions and lectures
http://www.youtube.com/onlineteaching
Chapter 10: Correlation and Regression
10.2: Regression
InstructionsPlease answer the following question in a minimum.docxdirkrplav
Instructions:
Please answer the following question in a minimum of 500 words. Be sure to include 2 citations.
Question:
On August 31, 2010, Chickasaw Industries issued $25 million of its 30-year, 6% convertible bonds dated August 31, priced to yield 5%. The bonds are convertible at the option of the investors into 1,500,000 shares of Chickasaw's common stock. Chickasaw records interest expense at the effective rate. On August 31, 2013, investors in Chickasaw's convertible bonds tendered 20% of the bonds for conversion into common stock that had a market value of $20 per share on the date of the conversion. On January 1, 2012, Chickasaw Industries issued $40 million of its 20-year, 7% bonds dated January 1 at a price to yield 8%. On December 31, 2013, the bonds were extinguished early through acquisition in the open market by Chickasaw for $40.5 million.
Required:
1.
Using the book value method, would recording the conversion of the 6% convertible bonds into common stock affect earnings? If so, by how much? Would earnings be affected if the market value method is used? If so, by how much?
2.
Were the 7% bonds issued at face value, at a discount, or at a premium? Explain.
3.
Would the amount of interest expense for the 7% bonds be higher in the first year or second year of the term to maturity? Explain.
4.
How should gain or loss on early extinguishment of debt be determined? Does the early extinguishment of the 7% bonds result in a gain or loss? Explain.
Statistics Questions to Answer.doc.rtf
2
*Note: An Excel Workbook has also been uploaded. Within that workbook are 8 XLS files which are included in 8 separate tabs. These files will be needed to answer most of the questions.This work is due Friday, September 19th
Q1)Fill in the blanks (show your work).
Variable
N
Mean
Median
TrMean
StDev
haircut
171
23.17
17.00
21.14
18.20
sleep
171
6.6477
7.0000
6.6487
0.8396
age
171
27.421
27.000
27.098
3.646
Correlations:haircut,sleep, age
haircut
sleep
sleep
-0.117
age
0.062
(1)
Covariances:haircut,sleep, age
haircut
sleep
age
haircut
(2)_
sleep
-1.79232
0.70491
age
4.12314
-0.45372
13.29226
Blank 1 =
Blank 2 =
Q2)Is the following statement correct? Explain why or why not.
“A correlation of 0 implies that no relationship exists between the two variables under study.”
Q3)Does how long children remain at the lunch table help predict how much they eat? The data in file lunchtime.xls (File is in Tab#1 of Excel Workbook) gives information on 20 toddlers observed over several months at a nursery school. “Time” is the average number of minutes a child spent at the table when lunch was served. “Calories” is the average number of calories the child consumed during lunch, calculated from careful observation of what the child ate each day.
Findthecorrelationforthesedata.
Supposeweweretorecordtimeatthetableinhoursratherthaninminutes.Howwouldthecorrelationchange?Why?
Writeasentenceortwoexplainingwhatthiscorrelationmeansfort.
The purpose of this tutorial is to show that Scilab can be considered as a powerful data mining tool, able to perform the widest possible range of important data mining tasks.
MIXTURES OF TRAINED REGRESSION CURVES MODELS FOR HANDWRITTEN ARABIC CHARACTER...gerogepatton
In this paper, we demonstrate how regression curves can be used to recognize 2D non-rigid handwritten shapes. Each shape is represented by a set of non-overlapping uniformly distributed landmarks. The underlying models utilize 2nd order of polynomials to model shapes within a training set. To estimate the regression models, we need to extract the required coefficients which describe the variations for a set of shape class. Hence, a least square method is used to estimate such modes. We proceed then, by training these coefficients using the apparatus Expectation Maximization algorithm. Recognition is carried out by finding the least error landmarks displacement with respect to the model curves. Handwritten isolated Arabic characters are used to evaluate our approach.
Bayesian Estimation for Missing Values in Latin Square Designinventionjournals
International Journal of Mathematics and Statistics Invention (IJMSI) is an international journal intended for professionals and researchers in all fields of computer science and electronics. IJMSI publishes research articles and reviews within the whole field Mathematics and Statistics, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online.
Numerical simulation of marangoni driven boundary layer flow over a flat plat...eSAT Journals
Abstract
A numerical algorithm is presented for studying Marangoni convection flow over a flat plate with an imposed temperature
distribution. Plate temperature varies with x in the following prescribed manner: where A and k are constants.
By means of similarity transformation, the original nonlinear partial differential equations of flow are transformed to a pair of
nonlinear ordinary differential equations. Subsequently they are reduced to a first order system and integrated using Newton
Raphson and adaptive Runge-Kutta methods. The computer codes are developed for this numerical analysis in Matlab
environment. Velocity profiles for various values of k, and temperature profiles for various Prandtl number and k are illustrated
graphically. The results of the present simulation are then compared with the previous works available in literature with good
agreement.
Keywords: Matlab, Marangoni Convection, Numerical Simulation, Surface Tension, Flat Plate.
This presentation covered the following topics :
1. Variance
2. Standard Deviation
3. Meaning and Types of Skewness
4. Related Examples
and is useful for B.Sc & M.Sc students.
Solving Transportation Problems with Hexagonal Fuzzy Numbers Using Best Candi...IJERA Editor
In this paper, we introduce a Fuzzy Transportation Problem (FTP) in which the values of transportation costs are
represented as hexagonal fuzzy numbers. We use the Best candidate method to solve the FTP. The Centroid
ranking technique is used to obtain the optimal solution.
Economics
Curve Fitting
macroeconomics
Curve fitting helps in capturing the trend in the data by assigning a single function
across the entire range.
If the functional relationship between the two quantities being graphed is known to be
within additive or multiplicative constants, it is common practice to transform the data in
such a way that the resulting line is a straight line.(by plotting) A process of quantitatively
estimating the trend of the outcomes, also known as regression or curve fitting, therefore
becomes necessary.
For a series of data, curve fitting is used to find the best fit curve. The produced equation is
used to find points anywhere along the curve. It also uses interpolation (exact fit to the data)
and smoothing.
Some people also refer it as regression analysis instead of curve fitting. The curve fitting
process fits equations of approximating curves to the raw field data. Nevertheless, for a
given set of data, the fitting curves of a given type are generally NOT unique.
Smoothing of the curve eliminates components like seasonal, cyclical and random
variations. Thus, a curve with a minimal deviation from all data points is desired. This
best-fitting curve can be obtained by the method of least squares.
What is curve fitting Curve fitting?
Curve fitting is the process of constructing a curve, or mathematical functions, which possess closest
proximity to the series of data. By the curve fitting we can mathematically construct the functional
relationship between the observed fact and parameter values, etc. It is highly effective in mathematical
modelling some natural processes.
What is a fitting model?
A fit model (sometimes fitting model) is a person who is used by a fashion designer or
clothing manufacturer to check the fit, drape and visual appearance of a design on a
'real' human being, effectively acting as a live mannequin.
What is a model fit statistics?
The goodness of fit of a statistical model describes how well it fits a set of
observations. Measures of goodness of fit typically summarize the discrepancy
between observed values and the values expected under the model in question.
What is a commercial model?
Commercial modeling is a more generalized type of modeling. There are high
fashion models, and then there are commercial models. ... They can model for
television, commercials, websites, magazines, newspapers, billboards and any other
type of advertisement. Most people who tell you they are models are “commercial”
models.
What is the exponential growth curve?
Growth of a system in which the amount being added to the system is proportional to the
amount already present: the bigger the system is, the greater the increase. ( See geometric
progression.) Note : In everyday speech, exponential growth means runaway expansion, such
as in population growth.
Why is population exponential?
Exponential population growth: When resources are unlimited, populations
exhibit exponential growth, resulting in a J-shaped curve.
Rajshahi Krishi Unnayan Bank is playing a vital role in the economic development of Bangladesh, especially in supporting farmers in sixteen districts of Rajshahi and Rangpur divisions. Agriculture is the foremost important part of the Bangladeshi economy.
Please Subscribe to this Channel for more solutions and lectures
http://www.youtube.com/onlineteaching
Chapter 10: Correlation and Regression
10.2: Regression
InstructionsPlease answer the following question in a minimum.docxdirkrplav
Instructions:
Please answer the following question in a minimum of 500 words. Be sure to include 2 citations.
Question:
On August 31, 2010, Chickasaw Industries issued $25 million of its 30-year, 6% convertible bonds dated August 31, priced to yield 5%. The bonds are convertible at the option of the investors into 1,500,000 shares of Chickasaw's common stock. Chickasaw records interest expense at the effective rate. On August 31, 2013, investors in Chickasaw's convertible bonds tendered 20% of the bonds for conversion into common stock that had a market value of $20 per share on the date of the conversion. On January 1, 2012, Chickasaw Industries issued $40 million of its 20-year, 7% bonds dated January 1 at a price to yield 8%. On December 31, 2013, the bonds were extinguished early through acquisition in the open market by Chickasaw for $40.5 million.
Required:
1.
Using the book value method, would recording the conversion of the 6% convertible bonds into common stock affect earnings? If so, by how much? Would earnings be affected if the market value method is used? If so, by how much?
2.
Were the 7% bonds issued at face value, at a discount, or at a premium? Explain.
3.
Would the amount of interest expense for the 7% bonds be higher in the first year or second year of the term to maturity? Explain.
4.
How should gain or loss on early extinguishment of debt be determined? Does the early extinguishment of the 7% bonds result in a gain or loss? Explain.
Statistics Questions to Answer.doc.rtf
2
*Note: An Excel Workbook has also been uploaded. Within that workbook are 8 XLS files which are included in 8 separate tabs. These files will be needed to answer most of the questions.This work is due Friday, September 19th
Q1)Fill in the blanks (show your work).
Variable
N
Mean
Median
TrMean
StDev
haircut
171
23.17
17.00
21.14
18.20
sleep
171
6.6477
7.0000
6.6487
0.8396
age
171
27.421
27.000
27.098
3.646
Correlations:haircut,sleep, age
haircut
sleep
sleep
-0.117
age
0.062
(1)
Covariances:haircut,sleep, age
haircut
sleep
age
haircut
(2)_
sleep
-1.79232
0.70491
age
4.12314
-0.45372
13.29226
Blank 1 =
Blank 2 =
Q2)Is the following statement correct? Explain why or why not.
“A correlation of 0 implies that no relationship exists between the two variables under study.”
Q3)Does how long children remain at the lunch table help predict how much they eat? The data in file lunchtime.xls (File is in Tab#1 of Excel Workbook) gives information on 20 toddlers observed over several months at a nursery school. “Time” is the average number of minutes a child spent at the table when lunch was served. “Calories” is the average number of calories the child consumed during lunch, calculated from careful observation of what the child ate each day.
Findthecorrelationforthesedata.
Supposeweweretorecordtimeatthetableinhoursratherthaninminutes.Howwouldthecorrelationchange?Why?
Writeasentenceortwoexplainingwhatthiscorrelationmeansfort.
The purpose of this tutorial is to show that Scilab can be considered as a powerful data mining tool, able to perform the widest possible range of important data mining tasks.
This is a book of probability and statisticsThis is a book of probability and statisticsThis is a book of probability and statisticsThis is a book of probability and statistics
TSTD 6251 Fall 2014SPSS Exercise and Assignment 120 PointsI.docxnanamonkton
TSTD 6251 Fall 2014
SPSS Exercise and Assignment 1
20 Points
In this class, we are going to study descriptive summary statistics and learn how to construct box plot. We are still working with univariate variable for this exercise.
Practice Example:
Admission receipts (in million of dollars) for a recent season are given below for the
n =
30 major league baseball teams:
19.4 26.6 22.9 44.5 24.4 19.0 27.5 19.9 22.8 19.0 16.9 15.2 25.7 19.0 15.5 17.1 15.6 10.6 16.2 15.6 15.4 18.2 15.5 14.2 9.5 9.9
10.7 11.9 26.7 17.5
Require:
a. Compute the mean, variance and standard deviation.
b. Find the sample median, first quartile, and third quartile.
c. Construct a boxplot and interpret the distribution of the data.
d. Discuss the distribution of this set of data by examining kurtosis and skewness
statistics, such as if the distribution is skewed to one side of the distribution, and if the
distribution shows a peaked/skinny curve or a spread out/flat curve.
SPSS Procedures for Computing Summary Statistics
:
Enter the 30 data values in the first column of SPSS
Data View
Tab
Variable View
and name this variable
receipts
Adjust
Decimals
to 3 decimal points
Type
Admission Receipts
($ mn)
in the
Label
column for output viewer
Return to
Data View
and click
A
nalyze
on the menu bar
Click the second menu
D
e
scriptive Statistics
Click
F
requencies …
Move
Admission Receipts
to the
Variable(s)
list by clicking the arrow button
Click
S
tatistics …
button at the top of the dialog box
Now, you can select the descriptive statistics according to what the question requires. For this practice question, it requires central tendency, dispersion, percentile and distribution statistics, so we click all the boxes
except for
P
ercentile(s): and Va
l
ues are group midpoints
.
Click
Continue
to return to the
Frequencies
dialog box
Click
OK
to generate descriptive statistic output which is pasted below:
The first table provides summary statistics and the second table lists frequencies, relative frequencies and cumulative frequencies. The statistics required for solving this problem are highlighted in red.
Statistics
Admission Receipts
N
Valid
30
Missing
0
Mean
18.76333
Std. Error of Mean
1.278590
Median
17.30000
Mode
19.000
Std. Deviation
7.003127
Variance
49.043782
Skewness
1.734
Std. Error of Skewness
.427
Kurtosis
5.160
Std. Error of Kurtosis
.833
Range
35.000
Minimum
9.500
Maximum
44.500
Sum
562.900
Percentiles
10
10.61000
20
14.40000
25
15.35000
30
15.50000
40
15.84000
50
17.30000
60
19.00000
70
19.75000
75
22.82500
80
24.10000
90
26.69000
Admission Receipts
Frequency
Percent
Valid Percent
Cumulative Percent
Valid
9.500
1
3.3
3.3
3.3
9.900
1
3.3
3.3
6.7
10.600
1
3.3
3.3
10.0
10.700
1
3.3
3.3
13.3
11.900
1
3.3
3.3
16.7
14.200
1
3.3
3.3
20.0
15.2.
Comparing the methods of Estimation of Three-Parameter Weibull distributionIOSRJM
Weibull distribution has many applications in engineering and plays an important role in reliability. Estimation of the location, scale and shape parameters of this distribution for both censored and non censored samples were considered by several authors. In this paper we compare Graphical oriented methods, “trial and error” approach, the approach of Jiang/Murthy and Maximum likelihood method developed by Bain & Engelhard for sample sets containing uncensored and censored sample. Importance of each method is discussed.
Finding the Extreme Values with some Application of Derivativesijtsrd
There are many different way of mathematics rules. Among them, we express finding the extreme values for the optimization problems that changes in the particle life with the derivatives. The derivative is the exact rate at which one quantity changes with respect to another. And them, we can compute the profit and loss of a process that a company or a system. Variety of optimization problems are solved by using derivatives. There were use derivatives to find the extreme values of functions, to determine and analyze the shape of graphs and to find numerically where a function equals zero. Kyi Sint | Kay Thi Win "Finding the Extreme Values with some Application of Derivatives" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-3 | Issue-6 , October 2019, URL: https://www.ijtsrd.com/papers/ijtsrd29347.pdf Paper URL: https://www.ijtsrd.com/mathemetics/other/29347/finding-the-extreme-values-with-some-application-of-derivatives/kyi-sint
A big task often faced by practitioners is in deciding the appropriate model to adopt
in fitting count datasets. This paper is aimed at investigating a suitable model for fitting
highly skewed count datasets. Among other models, COM-Poisson regression model was
proposed in this paper for fitting count data due to its varying normalizing constant. Some
statistical models were investigated along with the proposed model; these include
Poisson, Negative Binomial, Zero-Inflated, Zero-inflated Poisson and Quasi- Poisson
models. A real life dataset relating to visits to Doctor within a given period was equally
used to test the behavior of the underlying models. From the findings, it is recommended
that COM-Poisson regression model should be adopted in fitting highly skewed count
datasets irrespective of the type of dispersion.
1. Running head: TIME SERIES ANALYSIS OF GDP DATA 1
Time Series Analysis of Gross Domestic Product Data
Brian G. Wu
Oakland University
Competency Project for M.S. Applied Statistics
Advisor: Dorin Drignei
Committee Members: Gary McDonald, Theophilus Ogunyemi, and Xianggui Qu
April 1, 2016
2. TIME SERIES ANALYSIS OF GDP DATA 2
Abstract
The gross domestic product, or GDP, is a leading measurement of a nation’s production. This
report investigates the GDP of the United States of America, the United Kingdom, and Australia.
The GDP of all three countries increase exponentially, but deviations from the trend occur. This
report begins with a discussion of the definition including a comparison of Gross National
Product (GNP) and Gross Domestic Product (GDP) and the three methods of calculating the
GDP such as the production, expenditure, and income approaches. Following is a discussion of
the statistical methods used to analyze the GDP of the three countries. Using a time series
analysis, the nonstationary data with an exponential trend is differenced resulting in a more
stationary data. Then, a multivariate AR(2) model and three univariate AR(2) models are fitted to
the data followed by analysis of the residuals and predictions. With residuals passing the tests,
the report shows the predicted future GDP along with 95% prediction bounds for the next five
years. The results are obtained and compared using three statistical programs: ITSM, R, and
SAS.
Keywords: GDP, Time Series, Statistics
3. TIME SERIES ANALYSIS OF GDP DATA 3
Contents
Abstract............................................................................................................................... 2
Introduction......................................................................................................................... 4
A Statistical Analysis of the GDP ....................................................................................... 6
ARMA Model ................................................................................................................. 6
Predictions....................................................................................................................... 6
Results............................................................................................................................. 7
Conclusion ........................................................................................................................ 10
References......................................................................................................................... 12
Tables ................................................................................................................................ 13
Figures............................................................................................................................... 20
Codes................................................................................................................................. 41
R Code .......................................................................................................................... 41
SAS Code...................................................................................................................... 41
4. TIME SERIES ANALYSIS OF GDP DATA 4
Introduction
The Gross Domestic Product is one of the leading measurements of the economy. Ferrara
describes the GDP as “the most publicized and politically significant” (Indicator - Gross
Domestic Product, 2013). One of the countries with the largest economy is the United States of
America, which shifted from Gross National Product, or GNP, to GDP. While both GDP and
GNP measure in terms of produced goods and services, their coverage differs. According to the
Survey of Current Business article, the GDP “covers the goods and services produced by labor
and property located in the United States” while the GNP “covers the goods and services
produced by labor and property supplied by U.S. residents” (Gross Domestic Product as a
Measure of U.S. Production, 1991). In others words, GDP covers anything located within the
U.S. while GNP covers anything by U.S. citizens, regardless whether the citizens reside in the
U.S. or not.
There are three different ways to calculate the GDP, theoretically leading to the same
results, but experimentally leading to different results due to rounding error. The first is the
production approach. Ferrera describes it as “the final output, minus the value of intermediate
consumption” (Indicator - Gross Domestic Product, 2013). In the second approach, which is the
expenditure approach, 𝑌 = 𝐶 + 𝐼 + 𝐺 + (𝑋 − 𝑀), where 𝑌 is the GDP, 𝐶 is the consumption, 𝐼
is the investment, 𝐺 is the government expenditures, 𝑋 is the exports, 𝑀 is the imports. “To this
sum,” says Ferrera, referring to every household’s consumption of goods and services,
“statisticians add gross capital formation, which is a nation’s investment, minus its disposal of
assets to assist in capital formation or produce inventories of goods or valuables” (Indicator -
Gross Domestic Product, 2013). The final approach is the income approach. According to
Ferrera, this means summing up the compensation of employees, gross operating surplus, and
5. TIME SERIES ANALYSIS OF GDP DATA 5
mixed income as the subtotal, then subtracting any subsidies from productions or imports to
obtain the GDP (Indicator - Gross Domestic Product, 2013).
Table 1 displays the actual GDP (in millions of 2009 dollars) from 1920 through 2013 for
the United States (Johnston & Williamson, 2016), the United Kingdom (Williamson, 2016), and
Australia (Hutchinson, 2016), each obtained from MeasuringWorth. Figure 1, Figure 2, and
Figure 3 show a plot of the GDP for the United States, the United Kingdom, and Australia,
respectively, generated from Microsoft Excel. Notice that all three plots follow an exponential
trend. However, each of the three plots show a hump from 1940 through 1945. The GDP was
higher than expected during that time because of World War II. Around 2008, the United States
and the United Kingdom suffered a recession, which explains the dip in the plots. However, the
recession apparently did not affect Australia as much, as there does not seem to be a dip in the
plot for Australia’s GDP.
To analyze and predict the future GDP, time series analysis has been used. We compare
our analysis with three statistical packages. ITSM (Interactive Time Series Modeling) is a time
series software that comes with Introduction to Time Series and Forecasting by Brockwell and
Davis. R functions such as “ts”, “diff”, and “ar” are key functions used to create a time series
object, difference the data, and fit autoregressive models, respectively. SAS procedures such as
ARIMA and VARMAX easily fit autoregressive moving-average models to univariate and
multivariate time series data, respectively, along with predictions of future values. The methods
are further discussed in the following sections.
6. TIME SERIES ANALYSIS OF GDP DATA 6
A Statistical Analysis of the GDP
ARMA Model
{𝑋𝑡} is an ARMA(𝑝, 𝑞) process if {𝑋𝑡} is stationary and if for every 𝑡, 𝑋𝑡 − 𝜙1 𝑋𝑡−1 − ⋯ −
𝜙 𝑝 𝑋𝑡−𝑝 = 𝑍𝑡 + 𝜃1 𝑍𝑡−1 + ⋯ + 𝜃𝑞 𝑍𝑡−𝑞, where {𝑍𝑡}~𝑊𝑁(0, 𝜎2
), or 𝜙(𝐵)𝑋𝑡 = 𝜃(𝐵)𝑍𝑡,
{𝑍𝑡}~𝑊𝑁(0, 𝜎2
), where 𝜙(𝑧) ≔ 1 − 𝜙1 𝑧 − ⋯ − 𝜙 𝑝 𝑧 𝑝
and 𝜃(𝑧) ≔ 1 + 𝜃1 𝑧 + ⋯ + 𝜃 𝑞 𝑧 𝑞
are
polynomials and 𝐵 denotes the backward shift operator (Brockwell & Davis, 2002).
The univariate ARMA model can easily be extended to the multivariate case. {𝐗 𝑡} is an
ARMA(𝑝, 𝑞) process if {𝐗 𝑡} is stationary and if for every 𝑡, 𝐗 𝑡 − Φ1 𝐗 𝑡−1 − ⋯ − Φ 𝑝 𝐗 𝑡−𝑝 =
𝐙 𝑡 + Θ1 𝐙 𝑡−1 + ⋯ + Θ 𝑞 𝐙 𝑡−𝑞, where {𝐙 𝑡}~𝑊𝑁(𝟎, Σ), or Φ(𝐵)𝑿 𝑡 = Θ(𝐵)𝐙 𝑡, {𝐙 𝑡}~𝑊𝑁(𝟎, Σ),
where Φ(𝑧) ≔ 𝐼 − Φ1 𝑧 − ⋯ − Φ 𝑝 𝑧 𝑝
and Θ(𝑧) ≔ 𝐼 + Θ1 𝑧 + ⋯ + Θ 𝑞 𝑧 𝑞
are matrix-valued
polynomials, 𝐼 is the 𝑚 × 𝑚 identity matrix, and 𝐵 denotes the backward shift operator
(Brockwell & Davis, 2002).
The AICC is defined as AICC = −2 ln 𝐿(Φ1, … , Φ 𝑝, Σ) +
2(𝑝𝑚2+1)𝑚𝑛
𝑛𝑚−𝑝𝑚2−2
(Brockwell &
Davis, 2002). This measure allows us to choose the best model achieving minimum AICC among
a set of competing models.
Predictions
Predicting the future values is important. The best ℎ-step linear predictor is denoted by
𝑃𝑛 𝑋 𝑛+ℎ = 𝜇 + ∑ 𝑎𝑖(𝑋 𝑛+1−𝑖 − 𝜇)𝑛
𝑖=1 where 𝐚 𝑛 = (𝑎1, … , 𝑎 𝑛)′ satisfies Γ𝑛 𝐚 𝑛 = 𝜸 𝑛(ℎ), Γ𝑛 =
[𝛾(𝑖 − 𝑗)]𝑖,𝑗=1
𝑛
, and 𝜸 𝑛(ℎ) = (𝛾𝑛(ℎ), 𝛾𝑛(ℎ + 1), … , 𝛾𝑛(ℎ + 𝑛 − 1))′.
In the multivariate case, let 𝑃𝑛 𝐗 𝑛+ℎ be the best ℎ-step linear predictor based on all the
components of 𝐗1, … , 𝐗 𝑛. Going back to the multivariate AR(𝑝) process defined by 𝐗 𝑡 =
Φ1 𝐗 𝑡−1 + ⋯ + Φ 𝑝 𝐗 𝑡−𝑝 + 𝐙 𝑡, {𝐙 𝑡}~𝑊𝑁(𝟎, Σ), one applies the linear prediction operator 𝑃𝑛 to
7. TIME SERIES ANALYSIS OF GDP DATA 7
get 𝑃𝑛 𝐗 𝑛+ℎ = Φ1 𝑃𝑛 𝐗 𝑛+ℎ−1 + ⋯ + Φ 𝑝 𝑃𝑛 𝐗 𝑛+ℎ−𝑝 (Brockwell & Davis, 2002). The predictions
𝑃𝑛 𝐗 𝑛+ℎ are obtained recursively with respect to ℎ. Prediction formulas for undifferenced
multivariate time series are given in Brockwell and Davis page 252 (Brockwell & Davis, 2002).
To compute the prediction error covariance matrix, let 𝐗 𝑛+ℎ = ∑ Ψ𝑗 𝐙 𝑛+ℎ−𝑗
∞
𝑗=0 , where we
can compute Ψ𝑗 recursively from Ψ𝑗 = Θ𝑗 + ∑ Φ 𝑘Ψ𝑗−𝑘
∞
𝑘=1 , 𝑗 = 0,1, … with 𝑞 = 0. We define
Θ0 = 𝐼, Θ𝑗 = 0 for 𝑗 > 𝑞, Φ𝑗 = 0 for 𝑗 > 𝑝, and Ψ𝑗 = 0 for 𝑗 < 0. Then, for 𝑛 ≥ 𝑝, 𝑃𝑛 𝐗 𝑛+ℎ =
∑ Ψ𝑗 𝐙 𝑛+ℎ−𝑗
∞
𝑗=ℎ . Then, 𝐗 𝑛+ℎ − 𝑃𝑛 𝐗 𝑛+ℎ = ∑ Ψ𝑗 𝐙 𝑛+ℎ−𝑗
ℎ−1
𝑗=0 is the h-step prediction error with
covariance matrix 𝐸[(𝐗 𝑛+ℎ − 𝑃𝑛 𝐗 𝑛+ℎ)(𝐗 𝑛+ℎ − 𝑃𝑛 𝐗 𝑛+ℎ)′] = ∑ Ψ𝑗ΣΨ𝑗
′ℎ−1
𝑗=0 , 𝑛 ≥ 𝑝.
Results
To do a multivariate time series analysis of the three GDPs, each of the three series must have
equal number of observations and values between 1920 and 2013 were retained. Upon initial
glance in the original plots in Figures 1, 2, and 3, there seems to be a nonstationary trend. The
first step is to difference the data to render them stationary. In a stationary time series, the data
has constant mean and variance over time without any trend. The differenced data for the GDP
for the United States, the United Kingdom, and Australia are shown in Figures 4, 5, and 6,
respectively, obtained from R and plotted using Excel and they look more stationary i.e. constant
mean, variance, and correlation depends on distance between time points. Figure 7 shows the
sample auto- and cross-correlations 𝜌̂𝑖𝑗𝑘(ℎ), 𝑖, 𝑗, 𝑘 = 1,2,3, generated from ITSM. For each of
the autocorrelation plots, at least one of the autocorrelations at lags other than zero fall outside of
the upper bound, suggesting the data values are correlated to both within and between time
series.
For the GDP data, using the Yule-Walker method, I had ITSM generate an AR model of
order less than or equal to 10 with the least AICC. The result is an AR model of order two, which
8. TIME SERIES ANALYSIS OF GDP DATA 8
sets 𝑝 = 2 and 𝑞 = 0, giving the equation 𝐗 𝑡 = Φ1 𝐗 𝑡−1 + Φ2 𝐗2 + 𝐙 𝑡, {𝐙 𝑡}~𝑊𝑁(𝟎, Σ) with
parameters 𝜙0 = [
50234.8
6180.46
1865.64
], Φ̂1 = [
0.416947 0.931785 0.312629
0.027292 0.288190 0.170756
0.022081 0.076634 0.357414
],
Φ̂2 [
−0.149708 0.164620 3.144659
−0.004624 −0.190773 0.123329
−0.016030 0.077330 0.278963
], Σ̂ =
[
1.97495 × 1010
1.64800 × 109
3.90519 × 108
1.64800 × 109
2.90734 × 108
4.66759 × 107
3.90819 × 108
4.66759 × 107
7.10182 × 107
], with AICC = 6454.97. My results with
R, shown in Table 2, is identical to my results with ITSM, shown in Table 3, with AICC of
essentially zero, and different values for Σ̂. However, my results from SAS with AICC =
5849.45, shown in Table 4, has slightly different values from R and ITSM for all the parameters.
Using ITSM, the residual plots for the United States, and United Kingdom, and Australia
are shown in Series 1, 2, and 3, respectively, of Figure 8. The residuals do not seem to have an
apparent pattern, passing the white-noise assumption and making it feasible to make predictions.
At the same time, ITSM also plots the residuals autocorrelation function for each pairs of data,
shown in Figure 9. Since the residual vectors 𝐙̂ 𝑡 are nearly all within the confidence bounds, the
model seems to be a good fit. While this analysis works with ITSM and R with identical graphs,
there is no simple code for residual analysis yet for SAS. None of the three packages provide a
formal white-noise test for multivariate time series models.
Assume that the model for the GDP fitted to the multivariate series {𝐘𝑡, 𝑡 = 0, … ,93} is
correct, i.e., that Φ(𝐵)𝐗 𝑡 = 𝐙 𝑡, 𝐙 𝑡~WN(𝟎, Σ̂), where 𝐗 𝑡 = (1 − 𝐵)𝐘𝑡 −
(159570,15535,14364)′, 𝑡 = 1, … ,93, Φ(𝑧) ≔ 𝐼 − Φ̂1 𝐵 − Φ̂2 𝐵2
, and Φ̂1, Φ̂2, Σ̂ are the
matrices found in Table 4. From ITSM, the one- and two-step predictors of 𝐘94 and 𝐘95 are
𝑃̃93 𝐘94 = [
1.59 × 107
1.68 × 106
1.46 × 106
] and 𝑃̃93 𝐘95 = [
1.61 × 107
1.71 × 106
1.46 × 106
]. ITSM also easily calculates the predicted
9. TIME SERIES ANALYSIS OF GDP DATA 9
GDP from 2014 through 2023 along with its 95% prediction limits, shown in Figure 10. As
shown in Figures 11, 12, and 13 for the United States, the United Kingdom, and Australia,
respectively, the actual GDP is within the prediction limits from 2009 to 2013 with the exception
of the 2013 Australia GDP, which is above the upper limit.
SAS automatically gives the future GDP, its standard error, and its 95% prediction
bounds for years 2014 through 2023, shown in Table 5 and Figures 14, 15, and 16 for the United
States, the United Kingdom, and Australia, respectively. As it turns out, the 2014 GDP for U.S.
and U.K. of 15,961,700 and 1,705,000, fit within their respective bounds of (15,594,673,
16,166,143) and (1,651,906, 1,719,292). Assuming the model is correct, the GDP apparently
increases after 2015 while the prediction bounds widen year after year.
However, if I took out the last five data points (2009 through 2013) for all three
countries, refitted the model, and computed the predicted GDP, my actual data points are not all
within the predicted bounds. This is because the multivariate AR(2) model does not take into
account the 2008 recession when the data points are removed. This is shown in Figures 17, 18,
and 19 for the United States, the United Kingdom, and Australia, respectively.
I now compare my multivariate AR(2) results to my results for three separate univariate
AR(2) results using SAS. The differenced data for each of the three univariate data along with
the correlation plots, shown in Figures 20, 21, and 22 for the United States, the United Kingdom,
and Australia, respectively, are similar to the differenced data for the multivariate data. My
parameter estimates are shown in Table 6 with AIC of 2489.653, 2098.003, and 1969.965 for the
United States, the United Kingdom, and Australia, respectively.
The residual plots for the univariate models along with their corresponding correlation
plots obtained with SAS are shown in Figures 23, 24, and 25 for the United States, the United
10. TIME SERIES ANALYSIS OF GDP DATA 10
Kingdom, and Australia, respectively. The residuals look mostly uncorrelated and in the white-
noise probability plot in the bottom right-hand corner of each of the three figures, much of the
bars are above 0.05, making the models of adequate fit.
Assuming that the model for the GDP fitted to their respective univariate series is correct,
SAS easily calculates the future GDP, its standard error, and its 95% prediction bounds for years
2014 through 2023, shown in Table 7 and Figures 26, 27, and 28 for the United States, the
United Kingdom, and Australia, respectively. Because the univariate models do not take
correlation between the different countries’ GDP into account, the predictions reported by the
univariate models have wider bounds than those reported by the multivariate model.
Additionally, the multivariate model predicts higher future GDP than the univariate models do
for the United States and the United Kingdom, but the opposite for Australia. However, if I took
out the last five data points (2009 through 2013) for all three countries, refitted the model, and
computed the predicted GDP, I noticed that for all three countries, at least one point is out of
bounds, as shown in Figures 29, 30, and 31.
A potential line of further analysis is to difference twice each series. However, this
shortens further the sample size by one observation vector and the multivariate AR model has
many parameters to be estimated. The estimation becomes less precise with fewer data.
Conclusion
The multivariate AR(2) model better-fits the differenced GDP data of the United States, the
United Kingdom, and Australia from 1920 through 2013 than the separate univariate AR(2)
models do because the multivariate model takes into account the correlation between the
different countries’ GDP. While the residuals shown in the residual plots appear uncorrelated and
compatible with the hypothesis of white noise, the multivariate model reports substantially
11. TIME SERIES ANALYSIS OF GDP DATA 11
narrower prediction bounds than the univariate models do for the predicted 2014 through 2023
GDP. However, the plots show that the GDP for all three countries is expected to increase, but
with the prediction limits increasing for each successive prediction, it is difficult to make
predictions for the GDP further into the future. Additionally, variations from the model trend are
not all by chance, as assignable causes such as war or recessions can deviate the actual GDP
from the predicted GDP.
12. TIME SERIES ANALYSIS OF GDP DATA 12
References
Brockwell, P. J., & Davis, R. A. (2002). Introduction to Time Series and Forecasting. Springer
Science & Business Media.
Gross Domestic Product as a Measure of U.S. Production. (1991, August). Survey of Current
Business, 8. Retrieved February 26, 2016, from
http://www.bea.gov/scb/pdf/national/nipa/1991/0891od.pdf
Hutchinson, D. (2016). What Was the Australian GDP or CPI Then? Retrieved January 12, 2016,
from MeasuringWorth: www.measuringworth.com/datasets/australiadata/
Indicator - Gross Domestic Product. (2013). In M. H. Ferrara (Ed.), Gale Business Insights
Handbook of Investment Research (pp. 41-52). Detroit: Gale, Cengage Learning.
Retrieved February 24, 2016, from
http://go.galegroup.com.huaryu.kl.oakland.edu/ps/i.do?id=GALE|9781414499345&v=2.1
&u=lom_oaklandu&it=aboutBook&p=GVRL&sw=w
Johnston, L., & Williamson, S. H. (2016). What Was the U.S. GDP Then? Retrieved January 12,
2016, from MeasuringWorth: http://www.measuringworth.org/usgdp/
Williamson, S. H. (2016). What Was the U.K. GDP Then? Retrieved January 12, 2016, from
MeasuringWorth: http://www.measuringworth.com/ukgdp/
14. TIME SERIES ANALYSIS OF GDP DATA 14
Table 2
Multivariate AR Coefficient Estimates from R
Lag Variable U.S. U.K. Australia
0 U.S. 50,234.50 - -
U.K. - 6,180.46 -
Australia - - 1,865.64
1 U.S. 0.4170 0.9318 0.3126
U.K. 0.0273 0.2882 0.1708
Australia 0.0221 0.0766 0.3574
2 U.S. -0.1497 0.1646 0.3145
U.K. -0.0046 -0.1908 0.1233
Australia -0.0160 0.0773 0.2790
Variance U.S. 2.187E+10 1.825E+09 4.324E+08
U.K. 1.825E+09 3.219E+08 5.168E+07
Australia 4.324E+08 5.168E+07 7.863E+07
15. TIME SERIES ANALYSIS OF GDP DATA 15
Table 3
Multivariate AR Coefficient Estimates from ITSM
Lag Variable U.S. U.K. Australia
0 U.S. 50,234.80 - -
U.K. - 6,180.46 -
Australia - - 1,865.64
1 U.S. 0.4169 0.9318 0.3126
U.K. 0.0273 0.2882 0.1708
Australia 0.0221 0.0766 0.3574
2 U.S. -0.1497 0.1646 3.1447
U.K. -0.0046 -0.1908 0.1233
Australia -0.0160 0.0773 0.2790
White Noise
Covariance
Matrix
U.S. 1.975E+10 1.648E+09 3.905E+08
U.K. 1.648E+09 2.907E+08 4.668E+07
Australia 3.908E+08 4.668E+07 7.102E+07
16. TIME SERIES ANALYSIS OF GDP DATA 16
Table 4
Multivariate AR Coefficient Estimates from SAS
Lag Variable U.S. U.K. Australia
0 U.S. 50,774.48 - -
U.K. - 6,234.74 -
Australia - - 1,773.37
1 U.S. 0.4068 0.9999 0.2241
U.K. 0.0254 0.3211 0.1715
Australia 0.0209 0.0836 0.3489
2 U.S. -0.1381 -0.0249 3.5959
U.K. -0.0035 -0.2187 0.1564
Australia -0.0145 0.0536 0.3317
White Noise
Covariance
Matrix
U.S. 2.125E+10 1.713E+09 3.914E+08
U.K. 1.713E+09 2.955E+08 4.622E+07
Australia 3.914E+08 4.622E+07 7.476E+07
18. TIME SERIES ANALYSIS OF GDP DATA 18
Table 6
Univariate AR Coefficient Estimates from SAS
Country Lag Coefficient Variance
U.S. 1 0.65398 24,020,000,000
2 0.12905
Intercept 159,572.8
U.K. 1 0.66183 357,100,000
2 0.01682
Intercept 15,535.37
Australia 1 0.57638 89,023,459
2 0.34599
Intercept 14,364.09
25. TIME SERIES ANALYSIS OF GDP DATA 25
Figure 11. Multivariate AR(2) Predictions for U.S. GDP with Original Data Using ITSM
Figure 12. Multivariate AR(2) Predictions for U.K. GDP with Original Data Using ITSM
12,500,000
13,000,000
13,500,000
14,000,000
14,500,000
15,000,000
15,500,000
16,000,000
16,500,000
2009 2010 2011 2012 2013
United States
GDP Predicted GDP Lower Limit Upper Limit
1,350,000
1,400,000
1,450,000
1,500,000
1,550,000
1,600,000
1,650,000
1,700,000
1,750,000
2009 2010 2011 2012 2013
United Kingdom
GDP Predicted GDP Lower Limit Upper Limit
26. TIME SERIES ANALYSIS OF GDP DATA 26
Figure 13. Multivariate AR(2) Predictions for Australia GDP with Original Data Using ITSM
Figure 14. Multivariate AR(2) Predictions for U.S. GDP Using SAS
1,100,000
1,150,000
1,200,000
1,250,000
1,300,000
1,350,000
1,400,000
1,450,000
2009 2010 2011 2012 2013
Australia
GDP Predicted GDP Lower Limit Upper Limit
14,000,000
15,000,000
16,000,000
17,000,000
18,000,000
19,000,000
20,000,000
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
United States
Predicted GDP Lower Limit Upper Limit
27. TIME SERIES ANALYSIS OF GDP DATA 27
Figure 15. Multivariate AR(2) Predictions for U.K. GDP Using SAS
Figure 16. Multivariate AR(2) Predictions for Australia GDP Using SAS
1,400,000
1,500,000
1,600,000
1,700,000
1,800,000
1,900,000
2,000,000
2,100,000
2,200,000
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
United Kingdom
Predicted GDP Lower Limit Upper Limit
1,400,000
1,450,000
1,500,000
1,550,000
1,600,000
1,650,000
1,700,000
1,750,000
1,800,000
1,850,000
1,900,000
2014 2015 2016 2017 2018 2019 2020 2021 2022 2023
Australia
Predicted GDP Lower Limit Upper Limit
28. TIME SERIES ANALYSIS OF GDP DATA 28
Figure 17. Multivariate AR(2) Predictions for U.S. GDP with Original Data Using SAS
Figure 18. Multivariate AR(2) Predictions for U.K. GDP with Original Data Using SAS
13,000,000
13,500,000
14,000,000
14,500,000
15,000,000
15,500,000
16,000,000
16,500,000
17,000,000
17,500,000
2009 2010 2011 2012 2013
United States
GDP Predicted GDP Lower Limit Upper Limit
-
200,000
400,000
600,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2,000,000
2009 2010 2011 2012 2013
United Kingdom
GDP Predicted GDP Lower Limit Upper Limit
29. TIME SERIES ANALYSIS OF GDP DATA 29
Figure 19. Multivariate AR(2) Predictions for Australia GDP with Original Data Using SAS
1,150,000
1,200,000
1,250,000
1,300,000
1,350,000
1,400,000
1,450,000
1,500,000
2009 2010 2011 2012 2013
Australia
GDP Predicted GDP Lower Limit Upper Limit
30. TIME SERIES ANALYSIS OF GDP DATA 30
Figure 20. Differenced Data and Correlation Plots for U.S. GDP Using SAS
TrendandCorrelationAnalysisfor us(1)
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
IACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
PACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
ACF
-400000
-200000
0
200000
400000
600000
us(1)
0 20 40 60 80 100
Observation
31. TIME SERIES ANALYSIS OF GDP DATA 31
Figure 21. Differenced Data and Correlation Plots for U.K. GDP Using SAS
TrendandCorrelationAnalysisfor uk(1)
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
IACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
PACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
ACF
-75000
-50000
-25000
0
25000
50000
uk(1)
0 20 40 60 80 100
Observation
32. TIME SERIES ANALYSIS OF GDP DATA 32
Figure 22. Differenced Data and Correlation Plots for Australia GDP Using SAS
TrendandCorrelationAnalysisfor au(1)
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
IACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
PACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
ACF
0
20000
40000
au(1)
0 20 40 60 80 100
Observation
33. TIME SERIES ANALYSIS OF GDP DATA 33
Figure 23. Univariate AR(2) Residual Analysis of U.S. Data Using SAS
Residual CorrelationDiagnosticsfor us(1)
0 5 10 15
Lag
1.0
.05
.001
WhiteNoiseProb
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
IACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
PACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
ACF
34. TIME SERIES ANALYSIS OF GDP DATA 34
Figure 24. Univariate AR(2) Residual Analysis of U.K. Data Using SAS
Residual CorrelationDiagnosticsfor uk(1)
0 5 10 15
Lag
1.0
.05
.001
WhiteNoiseProb
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
IACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
PACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
ACF
35. TIME SERIES ANALYSIS OF GDP DATA 35
Figure 25. Univariate AR(2) Residual Analysis of Australia Data Using SAS
Residual CorrelationDiagnosticsfor au(1)
0 5 10 15
Lag
1.0
.05
.001
WhiteNoiseProb
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
IACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
PACF
0 5 10 15
Lag
-1.0
-0.5
0.0
0.5
1.0
ACF
36. TIME SERIES ANALYSIS OF GDP DATA 36
Figure 26. Univariate AR(2) Predictions for U.S. GDP Using SAS
14000000
16000000
18000000
20000000
Forecast
96 98 100 102 104
Obs
95% ConfidenceLimitsPredicted
Forecastsfor us
37. TIME SERIES ANALYSIS OF GDP DATA 37
Figure 27. Univariate AR(2) Predictions for U.K. GDP Using SAS
1400000
1600000
1800000
2000000
Forecast
96 98 100 102 104
Obs
95% ConfidenceLimitsPredicted
Forecastsfor uk
38. TIME SERIES ANALYSIS OF GDP DATA 38
Figure 28. Univariate AR(2) Predictions for Australia GDP Using SAS
1400000
1500000
1600000
1700000
1800000
1900000
Forecast
96 98 100 102 104
Obs
95% ConfidenceLimitsPredicted
Forecastsfor au
39. TIME SERIES ANALYSIS OF GDP DATA 39
Figure 29. Univariate AR(2) Predictions for U.S. GDP with Original Data Using SAS
Figure 30. Univariate AR(2) Predictions for U.K. GDP with Original Data Using SAS
-
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
2009 2010 2011 2012 2013
United States
GDP Predicted GDP Lower Limit Upper Limit
-
200,000
400,000
600,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2,000,000
2009 2010 2011 2012 2013
United Kingdom
GDP Predicted GDP Lower Limit Upper Limit
40. TIME SERIES ANALYSIS OF GDP DATA 40
Figure 31. Univariate AR(2) Predictions for Australia GDP with Original Data Using SAS
-
200,000
400,000
600,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
1,800,000
2009 2010 2011 2012 2013
Australia
GDP Predicted GDP Lower Limit Upper Limit
41. TIME SERIES ANALYSIS OF GDP DATA 41
Codes
R Code
us=read.table(“USgdp.txt”)
uk=read.table(“UKgdp.txt”)
gdp=matrix(NA,dim(us)[1],2)
gdp[,1]=us[,2]
gdp[,2]=uk[,2]
gdp[1:4,]
write(t(gdp),ncol=2,file=“gdp.tsm”)
au=read.table(“AUgdp.txt”)
gdp=matrix(NA,dim(us)[1]-1,3)
gdp[,1]=us[1:94,2]
gdp[,2]=uk[1:94,2]
gdp[,3]=au[,2]
write(t(gdp),ncol=3,file=“gdp.tsm”)
jj=ts(gdp)
plot(jj)
dljj=diff(jj,lag=1)
plot(dljj,type=“l”)
mod<-ar(dljj,se.fit=TRUE,n.ahead=10,order.max=10,dmean=T)
mod$x.mean - (mod$ar[1,,]%*%mod$x.mean + mod$ar[2,,]%*%mod$x.mean)
mod
mod$aic
42. TIME SERIES ANALYSIS OF GDP DATA 42
mod$resid
acf(mod$resid[3:93,1])
plot(mod$resid)
SAS Code
/* *** GDP for U.S., U.K., and Australia: Multivariate Time Series Analysis*** */
title1 'GDP';
data mseries;
input us uk au @@;
date=_n_;
datalines;
743030 210661 62220
725995 183182 70722
766310 192907 74444
867213 198622 76972
893916 207997 79980
914914 215316 85187
974698 208598 82692
984111 224641 86181
995390 226646 85387
1056600 233113 83749
966700 231258 84948
904800 220530 76972