initial 值的问题怎么解决

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

有没有同学在进行DSGE运算时，dynare的运算结果是这样？怎么样解决，不知道问题出在哪里了

POSTERIOR KERNEL OPTIMIZATION PROBLEM!
(minus) the hessian matrix at the "mode" is not positive definite!
=> posterior variance of the estimated parameters are not positive.
You should try to change the initial values of the parameters using
the estimated_params_init block, or use another optimization routine.
Warning: The results below are most likely wrong!
> In dynare_estimation_1 at 436
In dynare_estimation at 62
In gi at 643
In dynare at 132
Warning: Matrix is singular to working precision.
> In dynare_estimation_1 at 450
In dynare_estimation at 62
In gi at 643
In dynare at 132

RESULTS FROM POSTERIOR MAXIMIZATION
parameters
prior mean mode s.d. t-stat prior pstdev

gamma 0.870 0.8700 Inf 0.0000 gamm 0.5000
kappa 1.100 1.1000 Inf 0.0000 gamm 0.5000
h 0.700 0.7000 Inf 0.0000 beta 0.2000
s 5.000 5.0000 Inf 0.0000 gamm 0.2500
delta_2 0.700 0.7000 Inf 0.0000 gamm 0.5000
rho_b 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_l 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_a 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_i 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_z 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_tl 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_tk 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_tc 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_g 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_d 0.500 0.5000 Inf 0.0000 beta 0.2000
rho_A 0.500 0.5000 Inf 0.0000 beta 0.2000
sigma_b 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_l 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_a 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_i 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_z 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_tl 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_tk 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_tc 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_g 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_d 1.000 1.0000 Inf 0.0000 invg 4.0000
sigma_A 1.000 1.0000 Inf 0.0000 invg 4.0000
phi_z 0.200 0.2000 Inf 0.0000 gamm 0.1000
phi_tk 1.000 1.0000 Inf 0.0000 gamm 0.3000
phi_tl 0.500 0.5000 Inf 0.0000 gamm 0.2500
phi_g 0.070 0.0700 Inf 0.0000 gamm 0.3000
phi_kl 0.250 0.2500 Inf 0.0000 norm 0.1000
phi_kc 0.050 0.0500 Inf 0.0000 norm 0.1000
phi_lc 0.050 0.0500 Inf 0.0000 norm 0.1000
phi_d 0.050 0.0500 Inf 0.0000 norm 0.1000
gamma_z 0.150 0.1500 Inf 0.0000 gamm 0.2000
gamma_tk 0.150 0.1500 Inf 0.0000 gamm 0.2000
gamma_tl 0.150 0.1500 Inf 0.0000 gamm 0.2000
gamma_g 0.150 0.1500 Inf 0.0000 gamm 0.2000
gamma_d 0.150 0.1500 Inf 0.0000 gamm 0.2000

Log data density [Laplace approximation] is NaN.

MH: Multiple chains mode.
MH: Old _mh files successfully erased!
MH: Searching for initial values...
MH: I couldn't get a valid initial value in 100 trials.
MH: You should Reduce mh_init_scale...
MH: Parameter mh_init_scale is equal to 0.400000.
MH: Enter a new value...

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏2 回帖

关键词：Initial TIA Optimization maximization Successfully matrix positive initial another

posterior variance of the estimated parameters are not positive.
You should try to change the initial values of the parameters using
the estimated_params_init block, or use another optimization routine.

pls try "mode_compute = 6"
  http://www.dynare.org/DynareWiki/MonteCarloOptimization

谢谢epoh老师，还有一个问题我比较疑惑：数据一般要经过价格平减、季节调整、HP滤波，请问这三个步骤的顺序是怎么样的呢？ 有些论文是先进过季节调整，然后再平减，最后HP滤波，是不是这样的呢？

tawjuan 发表于 2012-11-20 10:05
谢谢epoh老师，还有一个问题我比较疑惑：数据一般要经过价格平减、季节调整、HP滤波，请问这三个步骤的顺序 ...

数据处理可以参考
An Estimated New Keynesian DSGE Model_Turkey.pdf
   An Estimated New Keynesian DSGE Model_Turkey.pdf (571.98 KB)

2012-11-20 20:12:36 上传

page 11/23    3.2 Data

利率可以参考
page 8/23
2.3.1 The Monetary Policy Rule

谢谢epoh老师的帮助！ 国外的期刊一般都是detrended\seasonal adjusted，但国内的处理方法千姿百态，经济学季刊上的处理方法是先进行季节调整，再HP滤波处理，所以很是纠结呢

tawjuan 发表于 2012-11-20 21:25
谢谢epoh老师的帮助！ 国外的期刊一般都是detrended\seasonal adjusted，但国内的处理方法千姿百态，经济学 ...

给你找份文献:
  RBCs and DSGEs: The Computational Approach to Business Cycle Theory and Evidence
     
data:data.xls  
     GDP:Real Gross Domestic Product
     Source: U.S. Department of Commerce: Bureau of Economic Analysis  
     Seasonal Adjustment:Seasonally Adjusted Annual Rate

%数据处理 data.m
[gdp] = xlsread('data.xls','B15:B203');
y=log(gdp);
[trend] = hpfilter(y,1600);
y=(y-trend);  %detrend

运行文件:model.mod
    replication_dp07_15.rar (13.12 KB)

2012-11-22 16:23:06 上传

请问楼主最终如何解决的？