Faculty of Science FRM 9649: Time Series Analysis & Forecasting

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

FACULTY OF SCIENCE

FRM 9649: TIME SERIES ANALYSIS & FORECASTING


TRIMESTER 1, 2020

ASSIGNMENT 2

NAME HAFENI TULONGENI HAMUKOTO


STUDENT NO 201055031
EMAIL tulongeni@live.com

LECTURER MORGAN KAMGA-PENE


DATE OF SUBMISSION 10 AUGUST 2020
DEADLINE 10 AUGUST 2020
QUESTION 1
1.1. TRUE
1.2. FALSE
1.3. TRUE
1.4. FLASE
1.5. TRUE

QUESTION 2
2.1.

Restuarant Sales
300

250

200
Sales

150

100

50

Month

2017 2018 2019

2.2.

Month 2017 2018 2019


Jan 1.078
Feb 0.917 0.96 1.055
Mar 0.948 1.003 1.109
Apr 0.805 0.955 1.062
May 0.948 0.991 1.115
Jun 0.806 0.943 1.063
Jul 0.95 1.013 1.063
Aug 0.957 1 1.12
Sep 0.776 0.958 1.027
Oct 1.021 0.982 1.061
Nov 1.018 1.045 1
Dec 1.07 1.078
TOTAL 10.216 10.928 11.753
MEAN 0.929 0.993 1.068 2.99
ADJUSTED 0.932 0.996 1.072 3
INDEX 93.2 99.6 107.2

2.3. No, there is no trend in the data.

Sales (Deseasonalized)
300

250

200
Sales

150

100

50

Months

2017 2018 2019

2.4.

Month 2017 2018 2019 2020F


January 259,6567 264,0562 263,0597 262,9219
February 252,1459 238,9558 237,8731 243,7134
March 248,927 247,992 247,2015 248,5672
April 190,9871 193,7751 191,2313 193,6937
May 197,4249 193,7751 195,8955 194,2849
June 150,2146 149,5984 149,2537 149,9187
July 155,5794 157,6305 154,8507 157,8734
August 163,0901 161,6466 162,3134 161,9055
September 118,0258 122,9839 117,5373 123,1467
October 139,485 131,0484 138,0597 131,5235
November 163,0901 167,6707 161,3806 168,2405
December 221,03 231,8548 219,2164 232,4593

(𝑌𝑌𝑌𝑌−𝑌𝑌𝑌𝑌−3)
2020F = Yt+1 = Yt-2+
3

QUESTION 3
3.1.

Net Worth
160

140

120

100
Worth

80

60

40

20

0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Year

3.2. Yes - Non-linear Trends: points clustered around or along a curve that is not a
straight line.
3.3. The time-series is not covariance-stationary as none of the figures are repeated.
3.4. By the Breusch-Pagan test (regressing squared residuals over the same set of
explanatory variables), we have X = 18R2 = 5,1720. Thus, according to the chi-squared
table, heteroskedasticity is not a serious concern.
3.5. By the D-W test, we have D-W = 0,1872. Thus, according to the Durbin-Watson
table, we should be concerned about positive serial correlation.
3.6.

Year t Worth(Yt) In Yt Zt Žt
1997 1 4,4 1,4816
1998 2 10,55 2,3561 0,8745
1999 3 23,2 3,1442 0,7880 0,6028
2000 4 38,7 3,6558 0,5117 0,5451
2001 5 44 3,7842 0,1284 0,3608
2002 6 49,5 3,9020 0,1178 0,1051
2003 7 55 4,0073 0,1054 0,0981
2004 8 68,5 4,2268 0,2195 0,0898
2005 9 88,5 4,4830 0,2562 0,1659
2006 10 89,8 4,4976 0,0146 0,1904
2007 11 88,05 4,4779 -0,0197 0,0292
2008 12 94,15 4,5449 0,0670 0,0064
2009 13 106 4,6634 0,1185 0,0642
2010 14 111,85 4,7172 0,0537 0,0986
2011 15 122,2 4,8057 0,0885 0,0553
2012 16 141,55 4,9527 0,1470 0,0785
2013 17 148,75 5,0023 0,0496 0,1176
2014 18 150,9 5,0166 0,0144 0,0526

3.7.
t=15; Žt = 0; In Yt-1 = 5.0166
ln Ýt = Žt + ln Yt-1 = 0+5.0166
Ýt = e5.0166 ≈ 150,9

You might also like