Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

ACADEMIC SESSION 2018/2019: SEMESTER II

EQC7006: Time Series Analysis

1.  Answer the following questions.

a)  Given below is an annual data for 12 years. Estimate trend-cycle component of the series using a single 3-period moving average (3 MA) and a 3-term weighted moving average (3 WMA) with weights of 0.25, 0.5, and 0.25.

(4 marks)

Write formulae used to estimate irregular component ( ) for an annual time series data (additive and multiplicative trend).

(1 mark) (Total: 5 marks)

2.  The following tables give estimated coefficients for three regression models. Y denotes the monthly temperature (in Celsius) series. T represents time trend (T = 1, 2, 3,  …,  192).  Di’s are seasonal dummies where  D1=1 for January and 0 otherwise, and D2, D3, D4, D5, D6, D7, D8, D9, D10, D11 and D12 are similarly defined for February, March until December, respectively.

Model 1

Dependent Variable: Y

Method: Least Squares

Sample: 2000M01 2015M12

Included observations: 192


Model 2

Dependent Variable: Y

Method: Least Squares

Sample: 2000M01 2015M12

Included observations: 192

a)  What estimates are given by the twelve coefficients of Model 1? (1 mark)

b)  In between Model 1 and Model 2, which one should be used? Why? (2 marks)

c)  Based on Model 1, perform a seasonal variation test at 5% significance level. State the null and alternative hypotheses of the test. (4 marks)

d)  Given  below  are  out-sample  (January  2016  –  December  2016)  forecast accuracy measures produced by Model 1. Interpret test set MAPE, ACF1 and Theil’s U values.

3.  The following time  series  plot  represents  monthly  data  of  rainfall  amount  (in millimeters) in Malaysia for the period of 16 years (January 2000 – December 2015).

a)  Evaluate and discuss the results of ADF and KPSS tests given below. Given that seasonal variation  is present, decide whether seasonal differencing is needed before any ARMA or ARIMA model can be estimated for the above series.

Null Hypothesis: RAINFALL has a unit root

Exogenous: Constant

Lag Length: 0 (Automatic - based on SIC, maxlag=14)

t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-8.939625

0.0000

Test critical values:

1% level

-3.464460

5% level

-2.876435

10% level

-2.574788

*MacKinnon (1996) one-sided p-values.

Null Hypothesis: RAINFALL has a unit root

Exogenous: Constant, Linear Trend

Lag Length: 0 (Automatic - based on SIC, maxlag=14)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -8.919751 0.0000

Test critical values:

1% level

5% level

10% level

-4.006566

-3.433401

-3.140550

*MacKinnon (1996) one-sided p-values.

Null Hypothesis: RAINFALL is stationary

Exogenous: Constant

Bandwidth: 1 (Newey-West automatic) using Bartlett kernel

LM-Stat.

Kwiatkowski-Phillips-Schmidt-Shin test statistic

0.155087

Asymptotic critical values*:                         1% level

0.739000

5% level

0.463000

10% level

0.347000

*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)

Null Hypothesis: RAINFALL is stationary

Exogenous: Constant, Linear Trend

Bandwidth: 1 (Newey-West automatic) using Bartlett kernel

LM-Stat.

Kwiatkowski-Phillips-Schmidt-Shin test statistic

0.135448

Asymptotic critical values*:                         1% level

0.216000

5% level

0.146000

10% level

0.119000

*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)

(2 marks)

b)  Inspect correlograms of the series given below. Based on your finding in (a), propose two ARMA  or  ARIMA  models  for  the  given  series  and  give justifications.

* Partial autocorrelations beyond lag 36 are all insignificant.

* Partial autocorrelations beyond lag 36 are all insignificant.

(5 marks)

c)  By looking at the time plot of the series, can you decide whether the proposed models should include a constant term or not?

(0.5 mark) (Total: 7.5 marks)

4.  The following time series plot represents monthly data of temperature (in Celcius) in Malaysia for the period of 16 years (January 2000 – December 2015).

a)  Evaluate  and  discuss  the  results  of  ADF  and  KPSS  tests  given  below. Assuming  that  seasonal  variation  is  present,  decide  whether  seasonal differencing is needed before any ARMA or ARIMA model can be estimated for the above series.

Null Hypothesis: TEMPERATURE has a unit root

Exogenous: Constant

Lag Length: 13 (Automatic - based on SIC, maxlag=14)

t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-2.683750

0.0787

Test critical values:

1% level

-3.464460

5% level

-2.876435

10% level

-2.574788

*MacKinnon (1996) one-sided p-values.

Null Hypothesis: TEMPERATURE has a unit root

Exogenous: Constant, Linear Trend

Lag Length: 13 (Automatic - based on SIC, maxlag=14)

t-Statistic Prob.*

Augmented Dickey-Fuller test statistic -2.714042 0.2322

Test critical values:

1% level

5% level

10% level

-4.006566

-3.433401

-3.140550

*MacKinnon (1996) one-sided p-values.

Null Hypothesis: TEMPERATURE is stationary

Exogenous: Constant

Bandwidth: 4 (Newey-West automatic) using Bartlett kernel

LM-Stat.

Kwiatkowski-Phillips-Schmidt-Shin test statistic

0.141894

Asymptotic critical values*:                         1% level

0.739000

5% level

0.463000

10% level

0.347000

*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1)

Null Hypothesis: TEMPERATURE is stationary

Exogenous: Constant, Linear Trend

Bandwidth: 4 (Newey-West automatic) using Bartlett kernel

LM-Stat.

Kwiatkowski-Phillips-Schmidt-Shin test statistic

0.130433

Asymptotic critical values*:                         1% level

0.216000

5% level

0.146000

10% level

0.119000

*Kwiatkowski-Phillips-Schmidt-Shin (1992, Table 1) (2 marks)

b)  Inspect correlograms of the series given below. Based on your finding in (a), propose two ARMA  or  ARIMA  models  for  the  given  series  and  give justifications.

* Partial autocorrelations beyond lag 36 are all insignificant.

* Partial autocorrelations beyond lag 36 are all insignificant. (5 marks)

c)  By looking at the time plot of the series, can you decide whether the proposed models should include a constant term or not?

(0.5 mark) (Total: 7.5 marks)

5.  Given below is estimation results of an ARIMA model for a monthly time series data, Y.

Dependent Variable: D(Y,0,12)

Method: ARMA Conditional Least Squares (Marquardt - EViews legacy)

Sample: 2000M01 2015M12

Included observations: 192

a)  Does the estimated process is covariance stationary? Explain. (1 mark)

b)  Does the estimated process is invertible? Explain. (1 mark)

c)  Write the model in terms of the backshift operator and without backshift operator. (3 marks)


d)  Table A below gives the last 24 observations of the series (January 2014 to December  2015).  Table  B  on  the  other  hand  gives  the  last  24  residuals generated by the model. Compute forecasts for January and February of 2016 and  95%  forecast  intervals  (given  that  forecast  error  variance:  January 2016=0.0663, February 2016=0.0754).