I have used STATA to estimate Seemingly unrelated bivariate ordered probit regression model. I realized that the ereturn list does not have information about either the mse or rmse. To compare this model with other models, I have to manually compute the estimates.
Could anyone kindly assist me with this task?
I first run the code below:
Code:
bioprobit (dnmcas=sq noncomply headon daytime weekday sideswipe nrintersec) (daccdttyp=nrintersec daytime noncomply othervio dens)
Then, this was the result of the estimation.
Code:
group(daccd | ttyp) | Freq. Percent Cum. ------------+----------------------------------- 1 | 285 19.18 19.18 2 | 1,201 80.82 100.00 ------------+----------------------------------- Total | 1,486 100.00 initial: log likelihood = -1134.8348 rescale: log likelihood = -1134.8348 rescale eq: log likelihood = -1121.0874 Iteration 0: log likelihood = -1121.0874 Iteration 1: log likelihood = -1121.0282 Iteration 2: log likelihood = -1121.0282 Seemingly unrelated bivariate ordered probit regression Number of obs = 1,486 Wald chi2(7) = 63.14 Log likelihood = -1121.0282 Prob > chi2 = 0.0000 ------------------------------------------------------------------------------ | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- dnmcas | sq | .1639291 .0896575 1.83 0.067 -.0117964 .3396546 noncomply | -.3144068 .0963615 -3.26 0.001 -.5032719 -.1255418 headon | -.3678404 .1746644 -2.11 0.035 -.7101764 -.0255044 daytime | -.4223297 .0824215 -5.12 0.000 -.5838729 -.2607864 weekday | -.2729407 .0881706 -3.10 0.002 -.4457519 -.1001295 sideswipe | -.2678797 .1526305 -1.76 0.079 -.56703 .0312706 nrintersec | -.2329192 .1024442 -2.27 0.023 -.4337061 -.0321323 -------------+---------------------------------------------------------------- daccdttyp | nrintersec | -.4297652 .0956313 -4.49 0.000 -.617199 -.2423313 daytime | .631744 .0954214 6.62 0.000 .4447215 .8187665 noncomply | -1.760099 .110548 -15.92 0.000 -1.976769 -1.543429 othervio | -1.47428 .3164721 -4.66 0.000 -2.094554 -.8540066 dens | .2381624 .1015682 2.34 0.019 .0390924 .4372323 -------------+---------------------------------------------------------------- athrho | _cons | .0881832 .0786048 1.12 0.262 -.0658793 .2422457 -------------+---------------------------------------------------------------- /cut11 | .5014163 .0976875 .3099524 .6928802 /cut12 | 1.506739 .1130826 1.285101 1.728377 /cut21 | -1.680589 .1074767 -1.891239 -1.469938 -------------+---------------------------------------------------------------- rho | .0879554 .0779967 -.0657841 .2376158 ------------------------------------------------------------------------------ LR test of indep. eqns. : chi2(1) = 1.27 Prob > chi2 = 0.2600
The scalars and macros of the estimation are as follows:
Code:
ereturn list scalars: e(rc) = 0 e(ll) = -1121.028167390545 e(converged) = 1 e(rank) = 16 e(k) = 16 e(k_eq) = 6 e(k_dv) = 2 e(ic) = 2 e(N) = 1486 e(k_eq_model) = 1 e(df_m) = 7 e(chi2) = 63.13531987062479 e(p) = 3.56044625833e-11 e(ll_0) = -1121.662573360585 e(k_aux) = 3 e(chi2_c) = 1.268811940080013 e(p_c) = .2599896774879261 macros: e(chi2_ct) : "LR" e(depvar) : "dnmcas daccdttyp" e(predict) : "bioprobit_p" e(cmd) : "bioprobit" e(chi2type) : "Wald" e(vce) : "oim" e(opt) : "ml" e(title) : "Seemingly unrelated bivariate ordered probit regression" e(ml_method) : "d2" e(user) : "bioprobit_d2" e(technique) : "nr" e(properties) : "b V"
All help and suggestions would be much appreciated. Thank you.
0 Response to How to estimate RMSE after bioprobit estimation?
Post a Comment