Department of statistics

Model Questions on Time Series analysis for 4th year Statistics students
Multiple Choose: Select the best answer

1. A set of quantitative observations arranged at a fixed time intervals is called:

A. Arithmetic series B. Discrete time series C. Harmonic series D. Continues time
series
2. Which one of the following is not the objective of time series analysis?
A. To identify the variability of the time series
B. To explain the variation of one time series variable by the variation of other time series
variable.
C. To predict future time series values
D. To control a physical system or business outcome.
E. None of the above
3. A time series consists of:
A. Short-term variations B. Trend C. Irregular variations D. All of the above
4. Identify the wrong statement.
A. Finding out the values for each of the components is called decomposition of a time
series.
B. Decomposition is done either by the Additive model or by the Multiplicative model of
analysis.
C. Seasonal variations are caused by infrequent occurrences such as wars, earthquakes,
floods etc.
D. The cyclical variations is composed of prosperity, recession, depression and recovery.
5. The general tendency of values of the data to increase or decrease during a long period of
time is called
A. Seasonal variation B. Trend C. Irregular fluctuations D. Cyclic
6. A time series component that contains the ups and downs within a year is called
A. Cyclic B. Random C. Seasonal variation D. Trend
7. Which one of the following is a time series data?
A. The daily closing prices over one year for a single financial security.
B. Today’s closing prices for each of the 500 companies
C. daily closing prices over one year for 500 companies
D. None of the above
8. Identify the wrong statement regarding on Additive Model.
A. It assumes the cyclic effects remain constant for all cycles
B. It is based on the assumption that the four components are interdependent.

1
 C. In this model the values of the four components are expressed in the original units
of measurement
D. It is appropriate when seasonal variations do not depend on the trend of the time
series.
9. Point out the test that is not used to check randomness.
A. Difference sign test B. Turning point Test C. Rank Test D. shapiro wilk test
10. Identify the wrong statement about turning point test.
A. It is used to find if a set of time series data is truly random.
B. The null hypothesis under the test is the data series are truly random.
C. It is a good test to uncover cyclicity.
D. The data series are truly random when the null hypothesis is rejected.
11. Point out the method that is not used to estimate trend.
A. Ratio-to-Trend Method
B. Least squares method
C. Weighted moving average method.
D. Exponential smoothing

Consider the following table (weekly output of a factory) to answer Questions 12, 13 and 14.

Week Actual output (in Forecast output Forecast errors

thousands)
1 50
2 53
3 56
4
12. Based on exponential smoothing technique ( = 0.6), the values of forecast output from
week 1 to 4 are respectively.
A. 58,61,64,67 C. 53,51.2,52.28,54.51
B. 50,54,56,60 D. 51,52.4,55.6,54.1
13. The Forecast errors respectively.
A. 0,-1,0, C.-1,1.6,-.04
B. -3,1.8,3.72 D. 8,8,8
14. What is the 2- week simple moving average for week 4?
A. 51.5 B. 53 C.54.5 D. 56
15. Identify the wrong statement about Simple Moving Average method.
A. If the number of period is increased, the series becomes much smoother.
B. The most recent observation receives the largest weight and the weights decrease for
older data values.
C. Moving averages vary less than the data values from which they are calculated.

2
 D. None of the above
16. Point out the method that is not used to estimate the seasonal indices.
A. Simple Average Method
B. Ratio to Moving Average Method
C. Ratio to trend
D. Pearson method
E. None
17. Which one of the following is not true about seasonal index?
A. It tells us what is the ratio between the time series value at a certain value and the
overall seasonal average.
B. If you have quarterly data then the sum of all seasonal indices should be 1 (100%).
C. Suppose the monthly production of a commodity for 3 years was given. If the Seasonal
Index for January production is 1.0488, then it indicates 4.8% above the annual average
production a commodity.
D. It can be used to smooth data by a process called deseasonalising.
18. Identify the wrong statement about seasonal component analysis.
A. The seasonal indices calculated using Ratio to Moving Average Method are free from trend,
cyclic and Irregular variations .
B. Ratio to moving average method is better because of its accuracy.
C. When we use Ratio to Trend Method, The first step is to obtain the trend values by the
method of least squares.
D. It is possible to estimate trend from deseasonalised data.
E. None
19. Identify the wrong expression.
(𝑤𝑒𝑟𝑒 ∇ is difference operator and B is Backward shift operator ).
A. ∇𝑌𝑡 = 𝑌𝑡 − 𝑌𝑡−1
B. ∇2𝑌𝑡 = 𝑌𝑡 + 𝑌𝑡−1
C. ∇2𝑌𝑡 = 𝑌𝑡 − 2𝑌𝑡−1 + 𝑌𝑡−2
D. 𝐵2𝑌𝑡 = 𝑌𝑡−2
20. A collection of random variables ordered in time is called a(n):
A. Stochastic process C. Non-Stationary Process
B. Stationary Process D. No
21. Which of the following is not the characteristics of weak stationary process:
A. the mean is constant
B. the covariance depends only on the time lag
C. the variance remain constant
D. the covariance depends only on location in time.
E. None

3
22. Identify the wrong statement.
A. Weak stationarity needs finite variance
B. If the distributions of the observations are normally distributed, the series is said to
possess strict stationarity.
C. Strict stationarity needs finite variance
D. The dependence of previous observations declines overtime is the main thought of
stationarity.
23. For a weak stationary process:
A. At lag k=2, 𝑐𝑜𝑣(𝑌1, 𝑌3) = 𝑐𝑜𝑣(𝑌2, 𝑌4) = 𝑐𝑜𝑣(𝑌3, 𝑌5) …
B. The auto covariance at lag k=0, is equal with variance of the random variable.
C. The autocorrelation at lag k=0, is equal to one.
D. None
24. Point out the wrong statement .
A. A useful plot for interpreting a set of autocorrelations coefficient is called a
correlogram
B. Correlogram is a very important diagnostic tool for the selection of a suitable model.
C. Correlogram is alternatively known as the sample autocorrelation function (acf).
D. None

Consider the following figure to answer Questions 24.

25. Identify the correct statement.

A. The plot is called correlogram.
B. The correlogram suggests that observations with smaller lag are positively correlated
C. Autocorrelation decreases as lag k increases.
D. It satisfies the stationary processes.
E. All of the above

4
26. A series of random shocks, normally and independently distributed around a mean of zero
with a constant variance but without serial correlation is called a
A. white noise process B. Random walk process without drift
B. Random walk process drift D. None
27. The correct form of Random walk process without drift is:
A. 𝑌𝑡 = 𝛽1 + 𝑌𝑡−1 + 𝑒𝑡 C. 𝑌𝑡 = 𝛽1 + 𝑌𝑡−1
B. 𝑌𝑡 = 𝑌𝑡−1 + 𝑒𝑡 D. 𝑌𝑡 = 𝛽1 + 𝑒𝑡
28. Which one of the following is a stationary process?
A. white noise process B. Random walk process without drift
B. Random walk process drift D. None
29. Which one the following is not used to detect stationarity?
A. Ljung–box test D. Augmented Dickey Fuller (ADF)test
B. KPSS test E. None
C. ACF plot
30. Nonstationary may not follow from the presence of
A. Outliers D. random walk
B. white noise E. trend
C. changing variance
31. non-stationary series can be made stationary through
A. Multiplication B. differencing C. Addition D. None

32. The value of a series at a current time period is a function of its immediately previous value
plus some error is called:
A. Autoregressive processes C. Stationary process
B. Moving average processes D. white nose process
33. Let 𝑌𝑡 be a stochastic process with mean 𝜇, then consider a model
𝑌𝑡 = 𝜇 + 𝜃𝑒𝑡−1 + 𝜃𝑒𝑡−2 + 𝑒𝑡, point out the wrong thought.
A. This model is called MA(2) C. For invertibility |𝜃| < 1
B. This model is called ARIMA(0,0,2 ) D. This model is called AR(2)

34. Identify the wrong statment

A. Stationarity of data is a fundamental requirement for time series analysis.
B. MA processes are always stationary
C. AR and ARMA processes are generally not stationary unless appropriate restrictions are
imposed on the model parameters.
D. AR models are called short memory models.

5
35. Which one of the following is not a correct statement?
A. For a stationary AR and ARMA; The autocorrelation function(acf) dies down gradually
as lag k increases
B. For MA; the acf cuts off after a certain lag k
C. For a Random walk; the acf remains at one for all values of lag k
D. Box–Jenkins (ARIMA) models require a complete time series, with equally spaced
temporal intervals.
E. When a series exhibits weak stationarity, Box-Jenkins analysis becomes infeasible.
36. ARIMA can be used for both seasonal and non-seasonal data.
A. True B. False
37. Which one of the following is used to identify ARIMA Processes?
A. Correlogram
B. Minimum information criteria(MINIC)
C. Extended Autocorrelation Function(ESAF)
D. Squared canonical correlations(SCAN)
E. All
38. Which one of the following a suitable ARIMA Processes, when there is an exponential
decay of negative spikes in the PACF and a significant negative spike at lag 1 in the ACF?
A. ARIMA(0,0,1) 𝜃1 > 0 C. ARIMA(1,0,0) 𝜙1 > 0
B. ARIMA(0,0,1) 𝜃1 < 0 D. ARIMA(1,0,0) 𝜙1 < 0
39. If more than 5% ACF spikes are outside ± 2⁄ , (where T is number of observations) then
√𝑇
the series is not white noise.
A. True B. False
40. Which of the following is not used to forecast future events?
A. Naïve approach C. Exponential smoothing
B. Moving Average D. None
C. ARIMA
41. Which of the following is not used to evaluate forecasting?
A. MSE
B. Mean absolute deviation
C. Mean absolute percentage error
D. None
42. Wheat crops badly damaged on account of rains is:
A. Cyclical movement B. trend C. Seasonal movement D. random movement
43. The second degree parabola is fitted to the time series when the variations are:
A. Linear B. Random C. Downward D. non-linear

6
 44. Smoothing parameter close to one gives more weight or influence to recent observations
over the forecast.
A. True B. False
45. What is the name of the process 𝑌𝑡 − 𝑌𝑡−1 = 𝑒𝑡 − 0.5𝑒𝑡−1
A. IMA(2,2) C. ARI(1,1)
B. IMA(1,1) D. None of the above
46. In an analysis, if you get
 The Dickey-Fuller unit root test for the series {𝑌𝑡} does not reject a unit root.
 The ACF for the series{𝑌𝑡} has a very, very slow decay.
 The PACF for the differences {∇𝑌 } has significant spikes at lags 1 and 2 (and is
negligible at higher lags).
Then which model is most consistent with these observations?
A. IMA(1,1) C. ARIMA(2,2,2)
B. ARI(2,1) D. IMA(2,2)
47. A trend is the better fitted trend for which the sum of squares of residuals is:
A. Maximum B. Positive C. Minimum D. Negative
48. Consider an invertible MA(2) process 𝑌𝑡 = 𝑒𝑡 − 𝜃1𝑒𝑡−1 − 𝜃2𝑒𝑡−2, Which statement is true?
A. Its PACF can decay exponentially
B. It is always stationary.
C. Its ACF is nonzero at lags k = 1 and k = 2 and is equal to zero when k > 2.
D. All of the above.

49. BIC penalizes overly-large models more strongly than AIC.

A. True B. False
50. One can use an automated algorithm in r-software to find the best ARIMA model by using:
A. Auto. Arima() C. ARIMA()
B. Arima() D. lm()

 B. Discrete time series

 D. To control a physical system or business outcome.
 D. All of the above
 C. Seasonal variations are caused by infrequent occurrences such as wars, earthquakes, floods
etc.
 B. Trend
 C. Seasonal variation
 A. The daily closing prices over one year for a single financial security.
 D. It is appropriate when seasonal variations do not depend on the trend of the time series.
 D. Shapiro Wilk test
 C. It is a good test to uncover cyclicity.
 A. Ratio-to-Trend Method
 B. 50,54,56,60
7
 A. 0,-1,0,
 C. 54.5
 A. If the number of period is increased, the series becomes much smoother.
 D. Pearson method
 C. Suppose the monthly production of a commodity for 3 years was given. If the Seasonal Index
for January production is 1.0488, then it indicates 4.8% above the annual average production a
commodity.
 B. Ratio to moving average method is better because of its accuracy.
 B. ∇2𝑌𝑡 = 𝑌𝑡 + 𝑌𝑡−1
 A. Stochastic process
 D. the covariance depends only on location in time.
 B. If the distributions of the observations are normally distributed, the series is said to possess
strict stationarity.
 B. The auto covariance at lag k=0, is equal with variance of the random variable.
 D. None
 E. All of the above
 A. white noise process
 B. 𝑌𝑡 = 𝑌𝑡−1 + 𝑒𝑡
 A. white noise process
 E. None
 B. white noise
 B. differencing
 B. Moving average processes
 B. This model is called ARIMA(0,0,2 )
 B. MA processes are always stationary
 E. When a series exhibits weak stationarity, Box-Jenkins analysis becomes infeasible.
 A. True
 E. All
 B. ARIMA(0,0,1) 𝜃1 < 0
 A. True
 D. None
 D. None
 B. trend
 D. non-linear
 A. True
 B. IMA(1,1)
 A. IMA(1,1)
 C. Minimum
 D. All of the above.
 A. True
 A. Auto.Arima()