Scribd
Upload
968 views

CMA Additive and Multiplicative

This document provides two examples of time series data and asks the reader to: 1. Calculate the centered moving average and seasonal indexes for the data. 2. Fit additive and multiplicative linear trend models to the centered moving average. 3. Explain how to determine which model is better for prediction. 4. Forecast values for the four quarters of the fifth year for each example. The document outlines the steps to preprocess time series data using a centered moving average, decompose the data into trend and seasonal components, fit linear trend models, and select the better predictive model to forecast future values.

Uploaded by

Shweta Niru
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
968 views

CMA Additive and Multiplicative

This document provides two examples of time series data and asks the reader to: 1. Calculate the centered moving average and seasonal indexes for the data. 2. Fit additive and multiplicative linear trend models to the centered moving average. 3. Explain how to determine which model is better for prediction. 4. Forecast values for the four quarters of the fifth year for each example. The document outlines the steps to preprocess time series data using a centered moving average, decompose the data into trend and seasonal components, fit linear trend models, and select the better predictive model to forecast future values.

Uploaded by

Shweta Niru
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Quantitative Finance II 2013 S2

Time Series Centered Moving Average (CMA)


Example 1: Consider the Quarterly sales data as shown in table.

Year
1

(i)
(ii)
(iii)

(iv)
(v)

Quarter
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4

Sales
(1000s)
4.8
4.1
6
6.5
5.8
5.2
6.8
7.4
6
5.6
7.5
7.8
6.3
5.9
8
8.4

Calculate the centered moving average (CMA) values for this time series.
Compute seasonal indexes for the 4-quarters based on the CMA.
Fit a linear trend to the CMA data for the
a. additive model:
= + + .
b. multiplicative model:
=
.
Explain how you can decide which one of Additive or Multiplicative model is better
model for prediction.
Forecast the financial data value for the four quarters of the 5th Year.

[Linear Regression Formula:

( )


( )

Solution:
By Assoc. Prof. (Dr) R. Boojhawon

Page 1

Quantitative Finance II 2013 S2

T
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Year
1

Sales
(1000s)
4.8
4.1
6
6.5
5.8
5.2
6.8
7.4
6
5.6
7.5
7.8
6.3
5.9
8
8.4

Quarter
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4

4quartely
MA

Centered
MA
Trend
(CMA)

Seasonal
Irregular value
under additive
model

Seasonal Irregular
value under
multiplicative
model

5.350
5.600
5.875
6.075
6.300
6.350
6.450
6.625
6.725
6.800
6.875
7.000
7.150

5.475
5.738
5.975
6.188
6.325
6.400
6.538
6.675
6.763
6.838
6.938
7.075

0.525
0.762
-0.175
-0.988
0.475
1.000
-0.538
-1.075
0.738
0.962
-0.638
-1.175

1.096
1.133
0.971
0.840
1.075
1.156
0.918
0.839
1.109
1.141
0.908
0.834

Additive Model
Seasonal Irreg
variation for
Quarter
1
2
3
4

y1

0.525
0.7625

y2
-0.175
-0.9875
0.475
1

y3
-0.5375
-1.075
0.7375
0.9625

y4
-0.6375
-1.175

sum =
Correction term
=

By Assoc. Prof. (Dr) R. Boojhawon

average=
seasonal index
-0.45
-1.079166667
0.579166667
0.908333333
-0.041666667
-0.010416667

Page 2

Effects =
Average-CT
-0.4395833
-1.06875
0.58958333
0.91875
0

Quantitative Finance II 2013 S2


Multiplicative
model

Seasonal Irreg
variation for
Quarter
1
2
3
4

yr1

y2
0.970711
0.840404
1.075099
1.15625

1.09589
1.132898

Additive
corrected
Seasonal
index
-0.440
-1.069
0.590
0.919
-0.440
-1.069
0.590
0.919
-0.440
-1.069
0.590
0.919
-0.440
-1.069
0.590
0.919

Multiplicative
corrected
Seasonal
index
0.931
0.836
1.092
1.141
0.931
0.836
1.092
1.141
0.931
0.836
1.092
1.141
0.931
0.836
1.092
1.141

y3
0.917782
0.8389513
1.1090573
1.1407678

average= seasonal
index
0.932200477
0.837759204
1.093348842
1.143305143

Effects
(Seasonal
Index) =
Average x
CT
0.93066171
0.83637633
1.09154407
1.1414179

Total of
means =

4.006613666

Correction
term (or adj
factor)=4/Total
of means

0.998349313

y4
0.908108108
0.833922261

Additive
adjusted
Deseasonalised
sales
5.240
5.169
5.410
5.581
6.240
6.269
6.210
6.481
6.440
6.669
6.910
6.881
6.740
6.969
7.410
7.481

Multiplicative
Deseasonalised
sales
5.158
4.902
5.497
5.695
6.232
6.217
6.230
6.483
6.447
6.696
6.871
6.834
6.769
7.054
7.329
7.359

Suppose we want to forecast, then we require to find the trend model equation
Let us use the SLR model
slope=
0.13461538
Intercept=
5.2661859

By Assoc. Prof. (Dr) R. Boojhawon

Page 3

Quantitative Finance II 2013 S2


We use the RMSE for both models based on the data set and the
corresponding estimates. The best model will be the one having least RMSE.

Forecast
Trend
5.4008013
5.5354167
5.6700321
5.8046474
5.9392628
6.0738782
6.2084936
6.343109
6.4777244
6.6123397
6.7469551
6.8815705
7.0161859
7.1508013
7.2854167
7.4200321

t
17
18
19
20

Year
5

Quarter
1
2
3
4

AM:
Forecast
=Forecast
Trend + SI
4.961
4.467
6.260
6.723
5.500
5.005
6.798
7.262
6.038
5.544
7.337
7.800
6.577
6.082
7.875
8.339

A SI
-0.43958
-1.06875
0.589583
0.91875

By Assoc. Prof. (Dr) R. Boojhawon

MM: Forecast
=Forecast
Trend * SI
5.026318934
4.629691453
6.189089834
6.625528506
5.527444468
5.080047936
6.776844331
7.240138146
6.028570002
5.530404419
7.364598828
7.854747787
6.529695536
5.980760902
7.952353324
8.469357427
RMSE=

M SI
0.9306617
0.8363763
1.0915441
1.1414179

A Error
-0.161
-0.367
-0.260
-0.223
0.300
0.195
0.002
0.138
-0.038
0.056
0.163
0.000
-0.277
-0.182
0.125
0.061
0.191

Trend
7.554647436
7.689262821
7.823878205
7.95849359

M Error
-0.22632
-0.52969
-0.18909
-0.12553
0.272556
0.119952
0.023156
0.159862
-0.02857
0.069596
0.135401
-0.05475
-0.2297
-0.08076
0.047647
-0.06936
0.192

A Forecast
7.115064103
6.620512821
8.413461538
8.87724359

M Forecast
7.03082107
6.43111739
8.54010782
9.08396707

Page 4

Quantitative Finance II 2013 S2


Example 2:

Consider the quarterly financial data sets


(I)
Year 1:

3.9 3.2 5.1 5.6

Year 3:

5.1 4.7 6.6 6.9

Year 2:
Year 4:

4.9 4.3 5.9 6.5


5.4 5.0 7.1 7.5

(II)

Year
1
2
3
4

1
7.0
15.0
23.2
31.3

Quarters
2
9.4
17.4
25.4
33.6

3
10.3
18.2
26.4
35.6

4
12.8
19.9
29.0
37.0

For each of these data sets:


(i)
(ii)
(iii)

(iv)
(v)

Calculate the centered moving average (CMA) values for this time series.
Compute seasonal indexes (additive/multiplicative) for the 4-quarters based on the CMA.
Fit a linear trend to the CMA data for the
a. additive model:
= + + .
b. multiplicative model:
=
.
Explain how you can decide which one of Additive or Multiplicative model is better
model for prediction.
Forecast the financial data value for the four quarters of the 5th Year.

[Linear Regression Formula:

By Assoc. Prof. (Dr) R. Boojhawon

( )

Page 5

You might also like