70.2K
Verified Solution
Link Copied!
Compare the MAE and RMSE of the first two questions. How would you evaluate your models in terms of forecasting performance?
AR Date: 05/09/22 Time: 21:02 Sample (adjusted): 5/02/2017 4/27/2022 Q-statistic probabilities adjusted for 5 AR MA terms Dependent Variable: LOG RETURN Method Least Squares Date: 05/09/22 Time: 20:58 Sample: 5/02/2017 4/27/2022 Included observations: 1257 Convergence achieved after 77 iterations Coefficient covariance computed using outer product of gradients Partial Correlation AC PAC Q-Stat Prob Q Autocorrelation 11 010 Variable Coefficient Std Error t-Statistic Prob. il AR(1) AR(6) AR(7) AR(8) AR(14) SIGMASQ 11 00 010 00 ill 1 0.002 0.002 0.0029 2 0.022 0.022 0.6244 3 0.001 0.001 0.6252 4 -0.031 -0.031 1.8172 5 0.032 0.032 3.1264 6 0.003 0.004 3.1367 0.077 7 -0.004 -0.005 3.1553 0.206 8 0.004 0.003 3.1747 0.365 9 0.015 0.017 3.4571 0.484 10 -0.010 -0.011 3.5799 0.611 11 0.010 0.009 3.7058 0.716 12 -0.010 -0.009 3.8365 0.798 13 -0.019 -0.018 4.2863 0.830 14 -0.015 -0.016 4.5771 0.870 15 -0.008 -0.006 4.6543 0.913 | 16 0.039 0.039 6.5624 0.833 17 0.028 0.027 7.5413 0.820 18 -0.039 -0.041 9.5119 0.733 19 0.017 0.017 9.8722 0.771 20 -0.049 -0.045 12.905 0.610 0.000402 0.000398 1.011054 0.3122 -0.015394 0.013276 -1.159551 0.2465 -0.159631 0.014046 - 11.36490 0.0000 0.152687 0.015851 9.632456 0.0000 -0.086839 0.017146 -5.064757 0.0000 -0.048459 0.023037 -2.103512 0.0356 0.000251 4.82E-06 52.08942 0.0000 0.060841 Mean dependent var 0.000399 0.056333 S.D. dependent var 0.016359 0.015892 Akaike info criterion -5.440177 0.315678 Schwarz criterion -5.41 1573 3426.151 Hannan-Quinn criter. -5.429427 13.49643 Durbin-Watson stat 1.995112 0.000000 01 00 R-squared Adjusted R-squared S.E. of regression Sum sauared resid Log likelihood F-statistic Prob(F-statistic) 1 Estimation Command: ==== EEEE LS(OPTMETHOD=OPG) LOG_RETURN C AR(1) AR(6) AR(7) AR(8) AR(14) Estimation Equation: ========= == LOG_RETURN = C(1) + [AR(1)=C(2),AR(6)=C(3),AR(7)=C(4), AR(8)=C(5),AR(14)=C(6), UNCOND] Substituted Coefficients: -------- LOG_RETURN = 0.000402275509989 + [AR(1=-0.0153941691692,AR(6)=-0.159631282895,AR(7)=0.152687443499, AR(8)=-0.0868389747629,AR(14)=- 0.0484592124461, UNCOND] MA Date: 05/09/22 Time: 21:26 Sample (adjusted): 5/02/2017 4/27/2022 Q-statistic probabilities adjusted for 8 ARMA terms Autocorrelation Partial Correlation AC PAC Q-Stat Prob 0.2288 Dependent Variable:LOG RETURN Method: ARMA Maximun Likelihood (OPG-BHHH) Date: 05/09/22 Tme: 21:12 Sample: 5/02/2017 4/27/2022 Included abservations: 1257 Convergence achieved after 68 terations Coefficient covariance computed using outer product of gradients Variable Coefficient Sid. Error Statistic Prob 0.000402 0.000409 0.984339 0.3251 MAX1) -0.018419 0.0 13574 -1 209843 MA2) 0.0 19497 0.013115 1.488844 0.1374 MA/G) -0.160626 0.0 15755 10.19507 0.0000 M447) 0.154717 0.015127 10.22782 0.0000 MA/S) -0.088707 0.015349 5.849238 0.0000 MA/13) -0.049728 0.021012 2.368838 0.0181 SIGMASO 0.000251 5.03E- 08 50.03517 0.0000 R-squared 0.059674 Mean dependent var 0.000399 Adjusted Rsquared 0.054404 S.D. dependent var 0.016359 S.E. al regression 0.015908 Akaikenla Akaikeinfo criterion 5.437354 Sum squared resid 0.316071 Schwarz criterion -5.404683 Log ikelihood 3425.377 HannanQuinn er 5.425088 F-statistic 11.32327 Durtain Watsonstat 1.996454 ProbF-statistic) 0.000000 1) 1 -0.005 -0.005 0.0343 2 0.008 0.008 0.0862 3 0.004 0.004 0.1093 4 -0.032 -0.032 1.4287 5 0.033 0.032 2.7671 6 0.000 0.001 2.7871 7 -0.004 -0.004 2.7879 8 -0.003 -0.005 2.8028 9 0.022 0.024 3.3975 0.065 | 10 -0.015 -0.016 3.7003 0.157 11 0.015 0.014 3.9849 0.263 12 0.013 0.0 13 4.1884 0.381 13 -0.012 -0.0 10 4.3844 0.498 14 -0.013 -0.016 4.5716 0.600 15 -0.030 -0.028 5.2830 0.577 16 0.045 0.048 8.2928 0.40 17 0.033 0.033 9.7084 0.375 18 -0.003 -0.004 9.7188 0.485 19 0.035 0.034 11.243 0.423 20 0.014 0.0 18 11.504 0.486 Inverted MA Roots 82 44+ 86 32+.881 .78 +.03 .69+.38 .13811 -55621 69-381 .13+ 8 11 55+.62 44.66 -32-681 -78-03 Estimation Command: -=-=- LS(OPTMETHOD=OPG) LOG_RETURN C MA(1) MA(2) MA(6) MA(7) MA(8) MA(13) MA(18) MA(20) Estimation Equation ========== LOG_RETURN = C(1) + [MA(1)=C(2),MA(2)=C(3), MA(6)=C(4),MA(7)=C(5), MA(8)=C(6),MA(13)=C(7), MA(18)=C(8),MA(20)=C(9), UNCOND,ESTSMPL="5/02/2017 4/27/2022"] Substituted Coefficients: ============ -======== LOG_RETURN = 0.000405134216902 + [MA (1)=-0.0014927413852, MA(2)=0.0115506432772, MA(6)=-0.164578034043,MA(7)=0.161453211999, MA(8)=- 0.0888531280931,MA(13)=-0.0561785369392,MA(18)=-0.0349606957725,MA(20)=-0.065883195185, UNCOND,ESTSMPL="5/02/2017 4/27/2022"] Ar & ma - Date: 05/09/22 Time: 21:40 Sample (adjusted): 5/02/2017 4/27/2022 Q-statistic probabilities adjusted for 6 ARMA tems Autocorrelation Partial Correlation AC PAC Q-Stat Prob Prob. li oli Dependent Variable: LOG RETURN Method: ARMA Maximum Likelihood (OPG-BHHH) Date: 05/09/22 Time: 21:39 Sample: 5/02/2017 4/27/2022 Included observations: 1257 Convergence achieved after 109 iterations Coefficient covariance computed using outer product of gradients Variable Coefficient Std. Error t-Statistic 0.000400 0.000438 0.913630 AR(1) -0.464334 0.084632 -5.486487 AR(6) -0.286741 0.103927 -2.759059 AR(7) -0.059696 0.116159 -0.513919 MA(1) 0.450634 0.088638 5.083961 MA(6) 0.135792 0.108455 1.252056 MA(7) 0.140771 0.108115 1.302047 SIGMASQ 0.000252 4.92E-06 51.11025 10 0.3611 0.0000 0.0059 0.6074 0.0000 0.2108 0.1931 0.0000 111 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.059351 Mean dependent var 0.054080 S.D. dependent var 0.015911 Akaike info criterion 0.316179 Schwarz criterion 3425.168 Hannan-Quinn criter 11.25819 Durbin-Watson stat 0.000000 0.000399 0.016359 -5.437022 -5.404332 -5.424736 1.994674 1 0.002 0.002 0.0040 2 0.015 0.015 0.3063 3 0.003 0.003 0.3148 4 0.032 0.032 1.5701 5 0.034 0.034 3.0292 6 0.004 -0.003 3.0485 70.004 0.005 3.0727 0.080 8 -0.014 0.015 3.3364 0.189 9 -0.015 -0.013 3.6332 0.304 10 0.005 0.004 3.6610 0.454 11 0.006 0.006 3.7065 0.592 12 0.026 0.026 4.5347 0.605 13 0.014 0.014 4.7999 0.684 14 0.048 0.046 7.7081 0.462 18 15 0.015 0.014 7.9959 0.535 100 101 16 0.041 0.041 10.162 0.426 17 0.030 0.031 11.329 0.416 18 0.034 0.038 12.784 0.385 19 0.013 0.015 13.017 0.447 20 0.041 0.038 15.141 0.369 Inverted AR Roots -.04+ 801 .67-401 -21 .67+ 371 - 58+ 481 .67+ 401 -.76+. 39 67-371 -58-.481 -04-.80i -76-39 .05-.761 -.73 Inverted MA Roots .05+.761 Estimation Command: LS(OPTMETHOD=OPG) LOG_RETURN C AR(1) MA(1) AR(6) MA(6) AR(7) MA(7) Estimation Equation: ========= LOG_RETURN = C(1) + [AR(1)=C(2),AR(6)=C(3), AR(7)=C(4), MA (1)=C(5),MA(6)=C(6),MA(7)=C(7), UNCOND, ESTSMPL="5/02/2017 4/27/2022"] = --========= Substituted Coefficients: ========= LOG_RETURN = 0.000399917415616 + [AR(1=-0.464334105486, AR(6)=-0.286741032667,AR(7)=- 0.0596964149171,MA(1)=0.450633593181,MA(6)=0.135791885676,MA(7)=0.140771140508, UNCOND,ESTSMPL="5/02/2017 4/27/2022"] AR Date: 05/09/22 Time: 21:02 Sample (adjusted): 5/02/2017 4/27/2022 Q-statistic probabilities adjusted for 5 AR MA terms Dependent Variable: LOG RETURN Method Least Squares Date: 05/09/22 Time: 20:58 Sample: 5/02/2017 4/27/2022 Included observations: 1257 Convergence achieved after 77 iterations Coefficient covariance computed using outer product of gradients Partial Correlation AC PAC Q-Stat Prob Q Autocorrelation 11 010 Variable Coefficient Std Error t-Statistic Prob. il AR(1) AR(6) AR(7) AR(8) AR(14) SIGMASQ 11 00 010 00 ill 1 0.002 0.002 0.0029 2 0.022 0.022 0.6244 3 0.001 0.001 0.6252 4 -0.031 -0.031 1.8172 5 0.032 0.032 3.1264 6 0.003 0.004 3.1367 0.077 7 -0.004 -0.005 3.1553 0.206 8 0.004 0.003 3.1747 0.365 9 0.015 0.017 3.4571 0.484 10 -0.010 -0.011 3.5799 0.611 11 0.010 0.009 3.7058 0.716 12 -0.010 -0.009 3.8365 0.798 13 -0.019 -0.018 4.2863 0.830 14 -0.015 -0.016 4.5771 0.870 15 -0.008 -0.006 4.6543 0.913 | 16 0.039 0.039 6.5624 0.833 17 0.028 0.027 7.5413 0.820 18 -0.039 -0.041 9.5119 0.733 19 0.017 0.017 9.8722 0.771 20 -0.049 -0.045 12.905 0.610 0.000402 0.000398 1.011054 0.3122 -0.015394 0.013276 -1.159551 0.2465 -0.159631 0.014046 - 11.36490 0.0000 0.152687 0.015851 9.632456 0.0000 -0.086839 0.017146 -5.064757 0.0000 -0.048459 0.023037 -2.103512 0.0356 0.000251 4.82E-06 52.08942 0.0000 0.060841 Mean dependent var 0.000399 0.056333 S.D. dependent var 0.016359 0.015892 Akaike info criterion -5.440177 0.315678 Schwarz criterion -5.41 1573 3426.151 Hannan-Quinn criter. -5.429427 13.49643 Durbin-Watson stat 1.995112 0.000000 01 00 R-squared Adjusted R-squared S.E. of regression Sum sauared resid Log likelihood F-statistic Prob(F-statistic) 1 Estimation Command: ==== EEEE LS(OPTMETHOD=OPG) LOG_RETURN C AR(1) AR(6) AR(7) AR(8) AR(14) Estimation Equation: ========= == LOG_RETURN = C(1) + [AR(1)=C(2),AR(6)=C(3),AR(7)=C(4), AR(8)=C(5),AR(14)=C(6), UNCOND] Substituted Coefficients: -------- LOG_RETURN = 0.000402275509989 + [AR(1=-0.0153941691692,AR(6)=-0.159631282895,AR(7)=0.152687443499, AR(8)=-0.0868389747629,AR(14)=- 0.0484592124461, UNCOND] MA Date: 05/09/22 Time: 21:26 Sample (adjusted): 5/02/2017 4/27/2022 Q-statistic probabilities adjusted for 8 ARMA terms Autocorrelation Partial Correlation AC PAC Q-Stat Prob 0.2288 Dependent Variable:LOG RETURN Method: ARMA Maximun Likelihood (OPG-BHHH) Date: 05/09/22 Tme: 21:12 Sample: 5/02/2017 4/27/2022 Included abservations: 1257 Convergence achieved after 68 terations Coefficient covariance computed using outer product of gradients Variable Coefficient Sid. Error Statistic Prob 0.000402 0.000409 0.984339 0.3251 MAX1) -0.018419 0.0 13574 -1 209843 MA2) 0.0 19497 0.013115 1.488844 0.1374 MA/G) -0.160626 0.0 15755 10.19507 0.0000 M447) 0.154717 0.015127 10.22782 0.0000 MA/S) -0.088707 0.015349 5.849238 0.0000 MA/13) -0.049728 0.021012 2.368838 0.0181 SIGMASO 0.000251 5.03E- 08 50.03517 0.0000 R-squared 0.059674 Mean dependent var 0.000399 Adjusted Rsquared 0.054404 S.D. dependent var 0.016359 S.E. al regression 0.015908 Akaikenla Akaikeinfo criterion 5.437354 Sum squared resid 0.316071 Schwarz criterion -5.404683 Log ikelihood 3425.377 HannanQuinn er 5.425088 F-statistic 11.32327 Durtain Watsonstat 1.996454 ProbF-statistic) 0.000000 1) 1 -0.005 -0.005 0.0343 2 0.008 0.008 0.0862 3 0.004 0.004 0.1093 4 -0.032 -0.032 1.4287 5 0.033 0.032 2.7671 6 0.000 0.001 2.7871 7 -0.004 -0.004 2.7879 8 -0.003 -0.005 2.8028 9 0.022 0.024 3.3975 0.065 | 10 -0.015 -0.016 3.7003 0.157 11 0.015 0.014 3.9849 0.263 12 0.013 0.0 13 4.1884 0.381 13 -0.012 -0.0 10 4.3844 0.498 14 -0.013 -0.016 4.5716 0.600 15 -0.030 -0.028 5.2830 0.577 16 0.045 0.048 8.2928 0.40 17 0.033 0.033 9.7084 0.375 18 -0.003 -0.004 9.7188 0.485 19 0.035 0.034 11.243 0.423 20 0.014 0.0 18 11.504 0.486 Inverted MA Roots 82 44+ 86 32+.881 .78 +.03 .69+.38 .13811 -55621 69-381 .13+ 8 11 55+.62 44.66 -32-681 -78-03 Estimation Command: -=-=- LS(OPTMETHOD=OPG) LOG_RETURN C MA(1) MA(2) MA(6) MA(7) MA(8) MA(13) MA(18) MA(20) Estimation Equation ========== LOG_RETURN = C(1) + [MA(1)=C(2),MA(2)=C(3), MA(6)=C(4),MA(7)=C(5), MA(8)=C(6),MA(13)=C(7), MA(18)=C(8),MA(20)=C(9), UNCOND,ESTSMPL="5/02/2017 4/27/2022"] Substituted Coefficients: ============ -======== LOG_RETURN = 0.000405134216902 + [MA (1)=-0.0014927413852, MA(2)=0.0115506432772, MA(6)=-0.164578034043,MA(7)=0.161453211999, MA(8)=- 0.0888531280931,MA(13)=-0.0561785369392,MA(18)=-0.0349606957725,MA(20)=-0.065883195185, UNCOND,ESTSMPL="5/02/2017 4/27/2022"] Ar & ma - Date: 05/09/22 Time: 21:40 Sample (adjusted): 5/02/2017 4/27/2022 Q-statistic probabilities adjusted for 6 ARMA tems Autocorrelation Partial Correlation AC PAC Q-Stat Prob Prob. li oli Dependent Variable: LOG RETURN Method: ARMA Maximum Likelihood (OPG-BHHH) Date: 05/09/22 Time: 21:39 Sample: 5/02/2017 4/27/2022 Included observations: 1257 Convergence achieved after 109 iterations Coefficient covariance computed using outer product of gradients Variable Coefficient Std. Error t-Statistic 0.000400 0.000438 0.913630 AR(1) -0.464334 0.084632 -5.486487 AR(6) -0.286741 0.103927 -2.759059 AR(7) -0.059696 0.116159 -0.513919 MA(1) 0.450634 0.088638 5.083961 MA(6) 0.135792 0.108455 1.252056 MA(7) 0.140771 0.108115 1.302047 SIGMASQ 0.000252 4.92E-06 51.11025 10 0.3611 0.0000 0.0059 0.6074 0.0000 0.2108 0.1931 0.0000 111 R-squared Adjusted R-squared S.E. of regression Sum squared resid Log likelihood F-statistic Prob(F-statistic) 0.059351 Mean dependent var 0.054080 S.D. dependent var 0.015911 Akaike info criterion 0.316179 Schwarz criterion 3425.168 Hannan-Quinn criter 11.25819 Durbin-Watson stat 0.000000 0.000399 0.016359 -5.437022 -5.404332 -5.424736 1.994674 1 0.002 0.002 0.0040 2 0.015 0.015 0.3063 3 0.003 0.003 0.3148 4 0.032 0.032 1.5701 5 0.034 0.034 3.0292 6 0.004 -0.003 3.0485 70.004 0.005 3.0727 0.080 8 -0.014 0.015 3.3364 0.189 9 -0.015 -0.013 3.6332 0.304 10 0.005 0.004 3.6610 0.454 11 0.006 0.006 3.7065 0.592 12 0.026 0.026 4.5347 0.605 13 0.014 0.014 4.7999 0.684 14 0.048 0.046 7.7081 0.462 18 15 0.015 0.014 7.9959 0.535 100 101 16 0.041 0.041 10.162 0.426 17 0.030 0.031 11.329 0.416 18 0.034 0.038 12.784 0.385 19 0.013 0.015 13.017 0.447 20 0.041 0.038 15.141 0.369 Inverted AR Roots -.04+ 801 .67-401 -21 .67+ 371 - 58+ 481 .67+ 401 -.76+. 39 67-371 -58-.481 -04-.80i -76-39 .05-.761 -.73 Inverted MA Roots .05+.761 Estimation Command: LS(OPTMETHOD=OPG) LOG_RETURN C AR(1) MA(1) AR(6) MA(6) AR(7) MA(7) Estimation Equation: ========= LOG_RETURN = C(1) + [AR(1)=C(2),AR(6)=C(3), AR(7)=C(4), MA (1)=C(5),MA(6)=C(6),MA(7)=C(7), UNCOND, ESTSMPL="5/02/2017 4/27/2022"] = --========= Substituted Coefficients: ========= LOG_RETURN = 0.000399917415616 + [AR(1=-0.464334105486, AR(6)=-0.286741032667,AR(7)=- 0.0596964149171,MA(1)=0.450633593181,MA(6)=0.135791885676,MA(7)=0.140771140508, UNCOND,ESTSMPL="5/02/2017 4/27/2022"]
Answer & Explanation
Solved by verified expert
