MULTIVARIATE ANALYSIS- TIME SERIES
Time: 1:30 hours Maximum marks: 40
PART – A
1. Answer any five questions out of the following: 5*2 = 10
a. In a straight-line equation Y = a + bX, b is the?
i. X-intercept
ii. Y-intercept
iii. Slope
iv. None of the above
b. Moving Average method is used for measurement of Trend when?
i. Trend is Linear
ii. Trend is Non-Linear
iii. Trend is Curvilinear
iv. None of the above
c. The most commonly used mathematical method for measuring the Trend is?
i. Moving Average Method
ii. Semi Average Method
iii. Method of Least-Squares
iv. None of the above
d. Sum of weights in Exponential Smoothing is?
i. <1
ii. >1
iii. 1
iv. None of the above
e. Seasonal Components in a Time Series
i. Cannot be predicted
ii. Are regular repeated patterns
iii. Are long runs of observations above or below the trend line
iv. Reflect shift in Time Series
f. If the actual demand for a period is 120 units but forecast demand was 110
units. The forecast error is?
i. -10
ii. +10
iii. -5
iv. +5
g. For an AR process to be considered stationary
i. The roots of the characteristic equation must lie inside the unit circle
ii. The roots of the characteristic equation must lie on the unit circle
iii. The roots of the characteristic equation must lie outside the unit circle
iv. The roots of the characteristic equation must be less than one in
absolute value
PART – B
2. Answer any four questions out of the following: 4*5 = 20
a. Why are different techniques employed to smooth a time series? Explain one
Smoothing Technique with a proper example to support your answer.
b. Explain Auto Regressive Models. What kind of plot can be used to determine
the order of AR models? Explain.
c. What is Stationarity in a time Series? Check if the following series is
stationary:
yt = 8yt-1 + 12yt-2 - 8yt-3 + εt
d. Apply Single Exponential Smoothing with α = 0.7 to forecast price based on
the following data:
Year Price
1782 112
1783 156
1784 121
1785 172
1786 187
1787 153
1788 152
1789 212
1790 135
1791 121
1792 263
1793 213
1794 161
1795 212
1796 223
1797 260
e. What do you mean by Auto Regressive Moving Average (ARMA model)?
How can we determine the order of an ARMA model?
f. What are the different methods to check for Stationarity? Explain in brief.
PART – C
3. Answer any one question: 1*10 = 10
a. Explain different types of Time Series Errors, their significance along with
their formulae.
b. Compute Mean Absolute Deviation (MAD), Mean Absolute Percentage Error
(MAPE), Mean Squared Deviation (MSD) based on the following time series
data:
Time
Serie Forecasted
Week
s Value
Value
1 54 52
2 43 46
3 49 51
4 47 48
5 39 43
6 40 38
7 48 52
8 57 61
9 52 57
10 47 51
11 38 36
c. Use Double Exponential Smoothing method with α = 0.2 and ß = 0.4 to
forecast production values for a factory. The values from year 1890 to 1910
are given as follows:
Yea Productio
r n
189 101
0 112
189 121
1 134
189 114
2 146
189 151
3 164
189 138
4 187
189 197
5 148
189
6
189
7
189
8
189
9
190
0
191
0
