Дипломная работа: Особенности инвестирования в криптовалюту

Внимание! Если размещение файла нарушает Ваши авторские права, то обязательно сообщите нам

Таблица № 3.7

Статистика при построении ARMA и GARCH-моделей выборки №2

Dependent Variable: Y

Method: ML - ARCH (Marquardt) - Normal distribution

Date: 03/30/19 Time: 20:07

Sample (adjusted): 11 500

Included observations: 490 after adjustments

Convergence achieved after 23 iterations

MA Backcast: 1 10

Presample variance: backcast (parameter = 0.7)

GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1)

Variable

Coefficient

Std. Error

z-Statistic

Prob.

AR(10)

-0.897740

0.011859

-75.70246

0.0000

MA(5)

0.044114

0.010391

4.245197

0.0000

MA(10)

0.961122

0.008188

117.3852

0.0000

Variance Equation

C

2.80E-05

8.12E-06

3.448743

0.0006

RESID(-1)^2

0.166957

0.022458

7.434321

0.0000

GARCH(-1)

0.836913

0.015817

52.91234

0.0000

R-squared

0.042983

Mean dependent var

0.004982

Adjusted R-squared

0.039053

S.D. dependent var

0.037723

S.E. of regression

0.036979

Akaike info criterion

-3.933614

Sum squared resid

0.665956

Schwarz criterion

-3.882254

Log likelihood

969.7353

Hannan-Quinn criter.

-3.913443

Durbin-Watson stat

1.881436

Модель ARMA(10;5,10) для выборки №2 имеет вид (3.6):

Модель GARCH(1;1) для выборки №2 имеет вид (3.7):

Рисунок 3.2 Волатильность доходности выборки №2 в %-годовых

Построение GARCH-модели для выборки №3.

Таблица № 3.8

Проверка на стационарность выборки №3 (тест Дики-Фуллера)

Null Hypothesis: Y has a unit root

Exogenous: Constant

Lag Length: 0 (Automatic - based on SIC, maxlag=16)

t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-18.19135

0.0000

Test critical values:

1% level

-3.448889

5% level

-2.869605

10% level

-2.571135

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(Y)

Method: Least Squares

Date: 04/06/19 Time: 18:17

Sample (adjusted): 2 350

Included observations: 349 after adjustments

Variable

Coefficient

Std. Error

t-Statistic

Prob.

Y(-1)

-0.975100

0.053602

-18.19135

0.0000

C

0.000919

0.002777

0.330783

0.7410

R-squared

0.488144

Mean dependent var

-0.000255

Adjusted R-squared

0.486669

S.D. dependent var

0.072400

S.E. of regression

0.051873

Akaike info criterion

-3.074341

Sum squared resid

0.933693

Schwarz criterion

-3.052249

Log likelihood

538.4725

Hannan-Quinn criter.

-3.065547

F-statistic

330.9251

Durbin-Watson stat

1.997986

Prob(F-statistic)

0.000000

Наблюдаемая статистика Дики-Фуллера меньше критических значений, что свидетельствует о стационарности ряда данных.

Таблица № 3.9 Коррелограмма доходностей выборки №3

Date: 04/06/19 Time: 18:18

Sample: 1 350

Included observations: 350

Autocorrelation

Partial Correlation

AC

PAC

Q-Stat

Prob

.|. |

.|. |

1

0.025

0.025

0.2183

0.640

.|. |

.|. |

2

0.051

0.050

1.1322

0.568

.|* |

.|* |

3

0.088

0.086

3.8651

0.276

.|. |

.|. |

4

0.068

0.063

5.5332

0.237

.|. |

.|. |

5

0.035

0.025

5.9808

0.308

.|. |

.|. |

6

0.037

0.023

6.4805

0.372

.|. |

.|. |

7

0.049

0.035

7.3484

0.394

.|. |

.|. |

8

0.036

0.023

7.8155

0.452

.|. |

.|. |

9

-0.021

-0.034

7.9702

0.537

.|. |

.|. |

10

0.005

-0.008

7.9796

0.631

.|. |

.|. |

11

-0.009

-0.018

8.0073

0.713

.|* |

.|* |

12

0.075

0.075

10.069

0.610

.|. |

.|. |

13

0.060

0.060

11.376

0.579

.|* |

.|. |

14

0.074

0.070

13.408

0.495

.|. |

.|. |

15

0.018

0.001

13.529

0.561

.|. |

.|. |

16

0.008

-0.015

13.552

0.632

*|. |

*|. |

17

-0.108

-0.134

17.899

0.395

.|. |

.|. |

18

0.002

-0.015

17.901

0.462

*|. |

*|. |

19

-0.071

-0.079

19.784

0.408

.|. |

.|. |

20

-0.024

-0.017

19.996

0.458

.|. |

.|. |

21

-0.013

0.004

20.064

0.517

.|. |

.|. |

22

-0.029

-0.005

20.375

0.560

.|. |

.|. |

23

0.017

0.050

20.485

0.613

.|. |

.|. |

24

-0.047

-0.030

21.335

0.619

.|. |

.|. |

25

-0.009

-0.002

21.368

0.672

.|. |

.|. |

26

0.021

0.005

21.530

0.714

.|. |

.|. |

27

0.048

0.046

22.411

0.716

.|. |

*|. |

28

-0.051

-0.067

23.417

0.712

.|. |

.|. |

29

0.016

0.029

23.517

0.752

.|. |

.|. |

30

0.046

0.055

24.317

0.758

*|. |

.|. |

31

-0.073

-0.044

26.373

0.703

.|. |

.|. |

32

-0.020

-0.002

26.524

0.740

.|. |

.|. |

33

0.021

0.028

26.691

0.773

.|. |

.|. |

34

-0.019

-0.024

26.830

0.804

.|. |

.|. |

35

-0.015

-0.013

26.919

0.834

.|. |

.|. |

36

-0.012

-0.022

26.979

0.862

Таблица № 3.10

Статистика при построении ARMA и GARCH-моделей выборки №3

Dependent Variable: Y

Method: ML - ARCH (Marquardt) - Normal distribution

Date: 03/30/19 Time: 20:35

Sample (adjusted): 20 350

Included observations: 331 after adjustments

Convergence achieved after 26 iterations

MA Backcast: 1 19

Presample variance: backcast (parameter = 0.7)

GARCH = C(4) + C(5)*RESID(-1)^2 + C(6)*GARCH(-1)

Variable

Coefficient

Std. Error

z-Statistic

Prob.

C

-0.003629

0.001327

-2.735792

0.0062

AR(19)

0.704550

0.028519

24.70423

0.0000

MA(19)

-0.960948

0.005736

-167.5249

0.0000

Variance Equation

C

5.76E-05

2.06E-05

2.790609

0.0053

RESID(-1)^2

0.107908

0.026869

4.016152

0.0001

GARCH(-1)

0.870394

0.029249

29.75811

0.0000

R-squared

0.160120

Mean dependent var

8.01E-05

Adjusted R-squared

0.154998

S.D. dependent var

0.052136

S.E. of regression

0.047926

Akaike info criterion

-3.383351

Sum squared resid

0.753369

Schwarz criterion

-3.314430

Log likelihood

565.9446

Hannan-Quinn criter.

-3.355862

Durbin-Watson stat

2.016732

Модель ARMA(19;19) для выборки №3 имеет вид (3.8):

Модель GARCH(1;1) для выборки №3 имеет вид (3.9):

Рисунок 3.3 Волатильность доходности выборки №3 в %-годовых

Построение GARCH-модели для выборки №4.

Таблица № 3.11

Проверка на стационарность выборки №4 (тест Дики-Фуллера)

Null Hypothesis: Y has a unit root

Exogenous: Constant

Lag Length: 0 (Automatic - based on SIC, maxlag=17)

t-Statistic

Prob.*

Augmented Dickey-Fuller test statistic

-19.19165

0.0000

Test critical values:

1% level

-3.443228

5% level

-2.867112

10% level

-2.569800

*MacKinnon (1996) one-sided p-values.

Augmented Dickey-Fuller Test Equation

Dependent Variable: D(Y)

Method: Least Squares

Date: 05/05/19 Time: 16:49

Sample (adjusted): 2 500

Included observations: 499 after adjustments

Variable

Coefficient

Std. Error

t-Statistic

Prob.

Y(-1)

-0.850862

0.044335

-19.19165

0.0000

C

0.000346

0.000273

1.269491

0.2049

R-squared

0.425646

Mean dependent var

-4.30E-06

Adjusted R-squared

0.424490

S.D. dependent var

0.008017

S.E. of regression

0.006082

Akaike info criterion

-7.363047

Sum squared resid

0.018383

Schwarz criterion

-7.346163

Log likelihood

1839.080

Hannan-Quinn criter.

-7.356421

F-statistic

368.3195

Durbin-Watson stat

2.008223

Prob(F-statistic)

0.000000

Наблюдаемая статистика Дики-Фуллера меньше критических значений, что свидетельствует о стационарности ряда данных.

Таблица № 3.12 Коррелограмма доходностей выборки №4

Date: 05/06/19 Time: 17:04

Sample: 1 500

Included observations: 500

Autocorrelation

Partial Correlation

AC

PAC

Q-Stat

Prob

.|* |

.|* |

1

0.149

0.149

11.180

0.001

.|. |

.|. |

2

0.074

0.053

13.912

0.001

.|. |

.|. |

3

-0.001

-0.020

13.912

0.003

.|. |

.|. |

4

-0.058

-0.061

15.614

0.004

.|. |

.|. |

5

-0.015

0.003

15.730

0.008

.|. |

.|. |

6

-0.002

0.008

15.731

0.015

.|. |

.|. |

7

0.002

0.001

15.734

0.028

.|. |

.|. |

8

0.021

0.017

15.955

0.043

*|. |

*|. |

9

-0.096

-0.105

20.642

0.014

*|. |

.|. |

10

-0.067

-0.043

22.948

0.011

.|. |

.|. |

11

-0.034

-0.005

23.547

0.015

*|. |

*|. |

12

-0.122

-0.112

31.220

0.002

.|. |

.|. |

13

-0.057

-0.036

32.882

0.002

.|. |

.|. |

14

-0.030

-0.012

33.361

0.003

.|. |

.|. |

15

-0.014

-0.008

33.462

0.004

.|. |

.|. |

16

0.057

0.051

35.141

0.004

.|. |

.|. |

17

0.022

0.007

35.397

0.006

.|. |

.|. |

18

-0.017

-0.040

35.551

0.008

.|. |

.|. |

19

-0.042

-0.048

36.457

0.009

.|. |

.|. |

20

-0.029

-0.007

36.900

0.012

.|. |

.|. |

21

-0.017

-0.026

37.047

0.017

.|. |

.|. |

22

-0.028

-0.046

37.445

0.021

.|. |

.|. |

23

0.014

0.010

37.551

0.028

.|. |

.|. |

24

-0.014

-0.039

37.651

0.038

.|. |

.|. |

25

-0.052

-0.056

39.080

0.036

.|. |

.|. |

26

0.006

0.024

39.100

0.048

.|. |

*|. |

27

-0.059

-0.067

40.919

0.042

.|. |

.|. |

28

-0.054

-0.049

42.458

0.039

.|. |

.|. |

29

-0.010

0.004

42.513

0.050

*|. |

*|. |

30

-0.077

-0.086

45.646

0.034

.|. |

.|. |

31

0.025

0.017

45.973

0.041

.|. |

*|. |

32

-0.065

-0.088

48.269

0.032

.|. |

.|. |

33

-0.032

-0.038

48.813

0.037

.|. |

.|. |

34

0.020

0.003

49.024

0.046

.|. |

.|. |

35

-0.041

-0.047

49.951

0.049

.|. |

.|. |

36

-0.032

-0.053

50.505

0.055

Таблица № 3.13

Статистика при построении ARMA и GARCH-моделей выборки №4

Dependent Variable: Y

Method: ML - ARCH

Date: 03/30/19 Time: 18:54

Sample (adjusted): 13 500

Included observations: 488 after adjustments

Convergence achieved after 42 iterations

MA Backcast: 4 12

Presample variance: backcast (parameter = 0.7)

GARCH = C(9) + C(10)*RESID(-1)^2 + C(11)*GARCH(-1)

Variable

Coefficient

Std. Error

z-Statistic

Prob.

C

0.000519

0.000109

4.766359

0.0000

AR(1)

0.581371

0.037356

15.56294

0.0000

AR(4)

-0.197749

0.053961

-3.664675

0.0002

AR(9)

0.335508

0.058027

5.781955

0.0000

AR(12)

-0.104271

0.029735

-3.506626

0.0005

MA(1)

-0.497902

0.026095

-19.08003

0.0000

MA(4)

0.150315

0.036092

4.164781

0.0000

MA(9)

-0.524412

0.040045

-13.09570

0.0000

Variance Equation

C

3.74E-06

1.58E-06

2.358862

0.0183

RESID(-1)^2

0.103658

0.037832

2.739943

0.0061

GARCH(-1)

0.789464

0.067419

11.70975

0.0000

R-squared

0.086942

Mean dependent var

0.000369

Adjusted R-squared

0.073626

S.D. dependent var

0.006166

S.E. of regression

0.005935

Akaike info criterion

-7.434911

Sum squared resid

0.016906

Schwarz criterion

-7.340457

Log likelihood

1825.118

Hannan-Quinn criter.

-7.397809

Durbin-Watson stat

1.932775

Модель ARMA(1,4,9,12;1,4,9) для выборки №4 имеет вид (3.10):

Источник: https://otherreferats.allbest.ru/download/1180544/