1. What is a time series?

A) A set of qualitative observations arranged randomly.

B) A set of quantitative observations arranged in chronological order.
C) A cross-sectional dataset with no time dimension.
D) A single observation recorded at a specific time.

Answer: B

2. Which of the following is NOT a component of a time series?

A) Trend
B) Seasonal
C) Cyclical
D) Random sampling

Answer: D

3. What does stationarity in a time series imply?

A) The mean and variance change over time.

B) The statistical properties (mean, variance, autocorrelation) remain constant over time.
C) The series has a deterministic trend.
D) The series is purely seasonal.

Answer: B

4. Which method is used to remove trend from a non-stationary time series?

A) Moving average
B) Exponential smoothing
C) Differencing
D) Seasonal adjustment

Answer: C

5. What is the purpose of the autocorrelation function (ACF)?

A) To measure the correlation between a series and its lagged values.

B) To decompose the series into trend and seasonal components.
C) To smooth the series using weighted averages.
D) To forecast future values directly.

Answer: A

6. In the ARIMA(p, d, q) model, what does "d" represent?

A) The number of autoregressive terms.

B) The number of moving average terms.
C) The number of differences needed for stationarity.
D) The seasonal period.
Answer: C

7. Which plot helps identify the order of an AR(p) model?

A) ACF plot
B) PACF plot
C) Time plot
D) Histogram

Answer: B

8. What is the Ljung-Box test used for?

A) To check if residuals are white noise (uncorrelated).

B) To estimate the parameters of an ARIMA model.
C) To identify the seasonal component.
D) To smooth the time series.

Answer: A

9. Which of the following is a characteristic of a moving average (MA) model?

A) It depends on past values of the series.

B) It depends on past forecast errors.
C) It requires differencing for stationarity.
D) It is non-linear.

Answer: B

10. What does the "I" in ARIMA stand for?

A) Independent
B) Integrated
C) Irregular
D) Inverted

Answer: B

11. Which method is used to estimate seasonal indices in a multiplicative model?

A) Simple averages
B) Ratio-to-moving average
C) Least squares regression
D) Exponential smoothing

Answer: B

12. What is the purpose of differencing in time series analysis?

A) To remove seasonality.
B) To stabilize the mean of the series (remove trend).
C) To smooth random fluctuations.
D) To decompose the series into components.
Answer: B

13. Which of the following is true about the Box-Jenkins methodology?

A) It only works for stationary series.

B) It involves model identification, estimation, and diagnostics.
C) It ignores autocorrelation in residuals.
D) It is limited to linear trends.

Answer: B

14. What does a significant spike at lag 1 in the ACF plot suggest?

A) A strong seasonal pattern.

B) A first-order autoregressive process (AR(1)).
C) A need for second differencing.
D) A deterministic trend.

Answer: B

15. Which of the following is NOT a step in time series forecasting?

A) Model identification
B) Parameter estimation
C) Hypothesis testing for causality
D) Model diagnostics

Answer: C

Multiple-Choice Questions (MCQs) on Time Series Analysis

1. What is the primary objective of time series analysis?

a) To predict future values based on historical data
b) To test hypotheses about population parameters
c) To determine the correlation between two variables
d) To classify data into different categories
Answer: a) To predict future values based on historical data

2. Which component of a time series is described as long-term

movement either upward or downward?
a) Seasonal
b) Cyclical
c) Trend
d) Irregular
Answer: c) Trend

3. A time series with constant mean and variance over time is

called:
a) Stationary
b) Nonstationary
c) Deterministic
d) Multiplicative
Answer: a) Stationary

4. Which method is used to remove non-stationarity in a time

series?
a) Smoothing
b) Differencing
c) Decomposition
d) Freehand method
Answer: b) Differencing

5. In the additive decomposition model, the components are

combined as:
a) Yt=Tt×St×Ct×ItYt=Tt×St×Ct×It
b) Yt=Tt+St+Ct+ItYt=Tt+St+Ct+It
c) Yt=Tt−St−Ct−ItYt=Tt−St−Ct−It
d) Yt=Tt/St/Ct/ItYt=Tt/St/Ct/It
Answer: b) Yt=Tt+St+Ct+ItYt=Tt+St+Ct+It

6. The Turning Points Test is used to:

a) Identify seasonal patterns
b) Test randomness in a time series
c) Estimate the trend component
d) Calculate autocorrelation
Answer: b) Test randomness in a time series
7. Which of the following is a limitation of the Freehand method
for trend estimation?
a) It is highly subjective
b) It assumes linearity
c) It ignores seasonal variations
d) It requires complex calculations
Answer: a) It is highly subjective

8. In exponential smoothing, if α=0.9α=0.9, it implies:

a) More weight is given to recent observations
b) Equal weight is given to all observations
c) More weight is given to older observations
d) The model ignores trends
Answer: a) More weight is given to recent observations

9. The Ratio to Moving Average method is primarily used to

estimate:
a) Trend component
b) Seasonal component
c) Cyclical component
d) Irregular component
Answer: b) Seasonal component

10. Which of the following is a characteristic of an AR(1) process?

a) Its PACF cuts off after lag 1
b) Its ACF cuts off after lag 1
c) Its PACF tails off exponentially
d) Its ACF tails off exponentially
Answer: a) Its PACF cuts off after lag 1

11. The Ljung-Box test is used for:

a) Checking residual autocorrelation
b) Estimating model parameters
c) Identifying seasonal patterns
d) Testing stationarity
Answer: a) Checking residual autocorrelation

12. A time series that exhibits periodic fluctuations with a fixed

interval (e.g., monthly sales peaks) is influenced by:
a) Trend
b) Seasonal component
c) Cyclical component
d) Irregular component
Answer: b) Seasonal component

13. In the Box-Jenkins methodology, the "I" in ARIMA stands for:

a) Integrated
b) Irregular
c) Independent
d) Inverted
Answer: a) Integrated

14. Which transformation is used to stabilize variance in a

multiplicative time series model?
a) Differencing
b) Logarithmic
c) Exponential smoothing
d) Centered moving average
Answer: b) Logarithmic

15. The autocorrelation function (ACF) of an MA(2) process will:

a) Cut off after lag 2
b) Tail off exponentially
c) Cut off after lag 1
d) Tail off sinusoidally
Answer: a) Cut off after lag 2

16. The Partial Autocorrelation Function (PACF) helps identify the

order of:
a) Moving Average (MA) process
b) Autoregressive (AR) process
c) Seasonal component
d) Cyclical component
Answer: b) Autoregressive (AR) process

17. Which method uses dummy variables to estimate seasonal

effects?
a) Ratio to Trend
b) Simple Average
c) Link Relative
d) Regression with seasonal dummies
Answer: d) Regression with seasonal dummies

18. A white noise process has:

a) Zero mean and constant variance
b) Time-dependent mean
c) Seasonal fluctuations
d) Autocorrelation at all lags
Answer: a) Zero mean and constant variance

19. In the equation Yt=ϕ1Yt−1+ϵtYt=ϕ1Yt−1+ϵt, the model is:

a) MA(1)
b) AR(1)
c) ARMA(1,1)
d) ARIMA(1,1,1)
Answer: b) AR(1)

20. The purpose of deseasonalization is to:

a) Remove the seasonal component from the data
b) Estimate the trend component
c) Smooth irregular fluctuations
d) Test for randomness
Answer: a) Remove the seasonal component from the data
1. Which component of a time series reflects long-term
upward/downward movements?
a) Seasonal
b) Cyclical
c) Trend
d) Irregular

2. A time series with observations that depend on previous values is

called:
a) Cross-sectional data
b) Serial-dependent data
c) Random noise
d) Stationary data

3. The multiplicative decomposition model is written as:

a) Yt=Tt+St+Ct+ItYt=Tt+St+Ct+It
b) Yt=Tt×St×Ct×ItYt=Tt×St×Ct×It
c) Yt=Tt×St+Ct+ItYt=Tt×St+Ct+It
d) Yt=Tt+St×Ct×ItYt=Tt+St×Ct×It

4. Which component accounts for unpredictable, short-term

fluctuations?
a) Trend
b) Seasonal
c) Cyclical
d) Irregular

Stationarity and Differencing

5. A stationary time series must have:

a) Constant mean and variance over time
 b) Increasing variance over time
c) Seasonal patterns
d) A linear trend

6. Differencing is primarily used to:

a) Estimate seasonal indices
b) Remove non-stationarity (e.g., trends)
c) Smooth irregular fluctuations
d) Calculate autocorrelation

7. The backward shift operator BB is defined such that:

a) BYt=Yt+1BYt=Yt+1
b) BYt=Yt−1BYt=Yt−1
c) BYt=Yt−Yt−1BYt=Yt−Yt−1
d) BYt=Yt+Yt−12BYt=2Yt+Yt−1

Trend and Seasonality Estimation

8. Which method assigns more weight to recent observations in

trend estimation?
a) Moving Average
b) Exponential Smoothing
c) Freehand Method
d) Semi-Average Method

9. The Ratio-to-Moving-Average method is used to estimate:

a) Trend
b) Seasonal indices
c) Cyclical variations
d) Irregular fluctuations
10. Deseasonalization in a multiplicative model involves:
a) Subtracting seasonal indices
b) Dividing by seasonal indices
c) Adding seasonal indices
d) Using dummy variables

Autocorrelation and Model Identification

11. The autocorrelation function (ACF) of an MA(1) process:

a) Cuts off after lag 1
b) Tails off exponentially
c) Shows seasonal spikes
d) Is always zero

12. A sharp cutoff in the PACF at lag 2 suggests the data follows
a(n):
a) MA(2) process
b) AR(2) process
c) ARMA(1,1) process
d) Random walk

13. The Ljung-Box test is used to:

a) Estimate model parameters
b) Check residual autocorrelation
c) Identify seasonal patterns
d) Test for stationarity
 ARIMA and Forecasting

14. In ARIMA(p, d, q), the parameter dd represents:

a) The order of the moving average component
b) The number of differencing steps
c) The number of autoregressive terms
d) The seasonal period

15. An ARIMA(1,1,1) model includes:

a) 1 autoregressive term, 1 differencing step, and 1 moving
average term
b) 1 seasonal component and 1 trend
c) 1 cyclical and 1 irregular component
d) A logarithmic transformation

16. A time series with a seasonal ARIMA model is denoted as:

a) ARIMA(p, d, q)
b) SARIMA(p, d, q)(P, D, Q)_s
c) ARMA(p, q)
d) ARCH(q)

Forecasting and Interpretation

17. If a forecast error has a mean close to zero, it indicates:

a) The forecast is unbiased
b) The model is overfitted
c) High volatility
d) Seasonality is present

18. The MSE (Mean Squared Error) measures:

a) Average squared difference between forecasts and actual
values
 b) The correlation between residuals
c) The trend strength
d) Seasonal variability

19. A forecast interval widens over time because:

a) Uncertainty increases with longer horizons
b) Seasonal effects dominate
c) The model is misspecified
d) Residuals are autocorrelated

Real-World Application

20. A company observes quarterly sales peaking every

December. Which component does this represent?
a) Trend
b) Seasonal
c) Cyclical
d) Irregular

Answer Key:

1. c | 2. b | 3. b | 4. d | 5. a | 6. b | 7. b | 8. b | 9. b | 10. b

2. a | 12. b | 13. b | 14. b | 15. a | 16. b | 17. a | 18. a | 19. a | 20. b

3. Which component is removed when deseasonalizing data in

a multiplicative model?
Answer: Divide by the seasonal index.
4. What does a slowly decaying ACF and a sharp cutoff in PACF
at lag 1 suggest?
Answer: The original series is non-stationary; the differenced series
is stationary.

5. Which model is indicated if PACF cuts off after lag 1 and ACF
decays exponentially?
Answer: AR(1).

6. Forecast using exponential smoothing (α=0.8, last

observation=120, previous forecast=110):
Answer: 118.

7. Ljung-Box test p-value=0.12 implies:

Answer: The model is adequate (residuals are white noise).

8. Variance of the error term in MA(1) with θ=0.5 and

Var(Yₜ)=2.25:
Answer: 1.8.

9. What remains after removing trend, seasonal, and cyclical

components in multiplicative decomposition?
Answer: Irregular component.

10. Correct order of Box-Jenkins methodology:

Answer: Identification, Estimation, Diagnostic Checking.

11. Autocorrelation at lag 3 for AR(1) with φ=0.7:

Answer: 0.34.

12. Differencing operator to remove seasonality in

quarterly data:
Answer: Yt−Yt−4Yt−Yt−4.
1. Which component of a time series is described as long-term
movement in the data?
a) Seasonal
b) Cyclical
c) Trend
d) Irregular
Answer: c) Trend
Explanation: The trend component represents the long-term upward or
downward movement in the data.

2. A multiplicative decomposition model is expressed as:

a) Yt=Tt+St+Ct+ItYt=Tt+St+Ct+It
b) Yt=Tt×St×Ct×ItYt=Tt×St×Ct×It
c) Yt=Tt×St+Ct+ItYt=Tt×St+Ct+It
d) Yt=Tt+St×Ct×ItYt=Tt+St×Ct×It
Answer: b) Yt=Tt×St×Ct×ItYt=Tt×St×Ct×It

3. Which test of randomness counts peaks and troughs in a series?

a) Difference Sign Test
b) Rank Test
c) Phase Length Test
d) Turning Points Test
Answer: d) Turning Points Test
4. The formula for the kk-th order autocorrelation coefficient (rkrk)
is:
a) ∑t=1n−k(xt−xˉ)(xt+k−xˉ)∑t=1n(xt−xˉ)2∑t=1n(xt−xˉ)2∑t=1n−k(xt−xˉ)(xt+k−xˉ)
b) ∑t=1n−kxtxt+k∑t=1nxt2∑t=1nxt2∑t=1n−kxtxt+k
c) ∑t=1n−k(xt−xˉ)∑t=1n(xt−xˉ)∑t=1n(xt−xˉ)∑t=1n−k(xt−xˉ)
d) ∑t=1n−kxt+k∑t=1nxt∑t=1nxt∑t=1n−kxt+k
Answer: a)

5. In exponential smoothing, if α=0.3α=0.3, the weight assigned to

the most recent observation is:
a) 0.3
b) 0.7
c) 0.09
d) 0.21
Answer: a) 0.3
Explanation: αα directly weights the most recent observation in
exponential smoothing.

6. A stationary time series must have:

a) Constant mean and variance, and autocovariance dependent on time
b) Constant mean, variance, and autocovariance independent of time
c) Zero mean and unit variance
d) Linearly increasing trend
Answer: b)
7. The PACF for an AR(2) model:
a) Cuts off after lag 2
b) Tails off gradually
c) Cuts off after lag 1
d) Is zero at all lags
Answer: a) Cuts off after lag 2

8. Differencing is used to:

a) Remove seasonality
b) Stabilize the variance
c) Remove non-stationarity in the mean
d) Smooth irregular fluctuations
Answer: c) Remove non-stationarity in the mean

9. In the Box-Jenkins methodology, the "I" in ARIMA stands for:

a) Independent
b) Integrated
c) Inverted
d) Irregular
Answer: b) Integrated

10. Which method is used to estimate seasonal indices by

averaging ratios to moving averages?
a) Link Relative Method
b) Ratio to Trend Method
c) Simple Average Method
d) Ratio to Moving Average Method
Answer: d) Ratio to Moving Average Method