Add tzDocumentation and activate .prn files in .gitignore

.gitignore

@@ -2,10 +2,11 @@
 .DS_Store
 
 # Use wildcards as well
+# *.prn
+
 *~
 *.mat
 *.pyc
-*.prn
 *.aux
 *.log
 *.dvi
@@ -19,6 +20,7 @@
 *.pdf
 *.npy
 *.o
+*.dll
 
 # Can also ignore all directories and files in a directory.
 # tmp/**/*

MatlabFiles/MSV/readme_msv.prn

+**********************************************************************
+************************** Important Notes ***************************
+**********************************************************************
+When eu=[0,1] in the original gensys setup, it means that in our MSV setting that exogenous persistence is put
+  in a wrong place when we reorder the roots.  Thus, there is no such [0,1] return in our MSV solution.
+
+When eu=[1,0] in the original gensys setup, it means that the MSV solution must exist with another order.  See
+  example3_10_sw.m in D:\ZhaData\WorkDisk\LiuWZ\ZhaNotes\UnderstandingMSV\OneDimentionCaseImportant.
+
+With our MSV solution, we abuse the notation eu=[1,0], which really means that we have multiple MSV solutions.
+  But multiple MSV solutions can occur only in complex numbers.  The imaginary part of the solution can be scaled
+  arbitrarily, which gives sunspot solutions.  Thus, the real MSV solution will be unique.
+
+If we encourter eu=[3,?], it means that all MSV solutions are explosive.  Thus, no stable MSV solution.
+
+
+
+========================= New files for MSV solutions ==============================
+msv_complex_AR.m: gives an MSV solution (upon the order that first works) with a flag indicating if [1,0].  If
+                    [1;0], the complex solution exits and thus the MSV solution is NOT unique.
+                    AR: autoregressive or gensys form;
+msv_all_complex_AR.m: gives all MSV-like solutions (including complex solutions).  That is, all the solutions that
+                        come from the undetermined coefficients method.
+
+
+========================= Old files for MSV solutions ==============================
+msv_simple.m:  works only for the first order if an msv solution (even if it is non-unique where the solution is
+                 complex) exist.  It does not try the second order (where the exogenous persistence may be larger
+                 than the endogenous persistence).
+
+msv_one.m: works only for the unqiue MSV solution (thus, no complex solution) for the order that comes first.  If no
+             solution is found for all orders, then no unique MSV solution (but there may be multiple MSV solutions
+             such as a complex solution).
+
+msv_all.m:  gives all MSV-like real solutions (thus, some solutions are NOT an MSV solution).

tzDocumentation/Template_datainp_Markov.prn

+/******************************************************************************/
+/********************* Markov State Variable Information **********************/
+/******************************************************************************/
+
+//== Flat Independent Markov States and Simple Restrictions ==//
+
+
+//1st state variable: learning gain.
+//2nd state variable: average seigniorage.
+//3rd state variable: seigniorage shocks.
+//4th state variable: after-escape shocks.
+//+
+//Case: tv_varibles == (TV_GAIN | TV_AD | TV_DSHOCKS) && syntypes == S_D_SHOCKS.  See datainp_setup.prn.
+//== Number Independent State Variables ==//
+2
+
+//=====================================================//
+//== state_variable[i] (1 <= i <= n_state_variables) ==//
+//=====================================================//
+//== Number of states for state_variable[1] ==//
+2
+
+
+//----------------- Prior for the transition matrix Q_k -----------------
+//== Each column contains the parameters for a Dirichlet prior on the corresponding
+//== column of the transition matrix.  Each element must be positive.  For each column,
+//== the relative size of the prior elements determine the relative size of the elements
+//== of the transition matrix and overall larger sizes implies a tighter prior.
+//If it is an identity matrix, we have no prior (i.e., a flat prior in this case).
+//Examples:
+//2.000000 1.000000
+//1.000000 2.000000
+//2.000000 1.000000 1.000000
+//1.000000 2.000000 1.000000
+//1.000000 1.000000 2.000000
+//== Transition matrix prior for state_variable[1]. (n_states x n_states) ==//
+1.000000 1.000000
+1.000000 1.000000
+
+
+//----------------- The semi-free parameter vector free -----------------
+//The vector, free, indictates the number of semi-free parameters in each column of Q_k.  By semi-free, we
+//  meen that the sum of each column in Q_k is 1.0 so that the number of true free parameters is one less.
+//== Free Dirichet dimensions for state_variable[1]  ==//
+2 2
+
+
+//----------------- The restriction matrix R_j -----------------
+//== The jth restriction matrix is n_states x free[j].  Each row of the restriction
+//== matrix has exactly one non-zero entry and the sum of each column must be one.
+//Example 1:
+//1 0 0
+//0 1 0
+//0 0 1
+//Example 2 (3 states but 2 semi-free parameters for column j of Q_1):
+//0 0
+//1 0
+//0 1
+//which gives the free-parameter vector [0 q_2 q_3] where q_2 + q_3 = 1.0.
+//== Column restrictions for state_variable[1] ==//
+1 0
+0 1
+
+1 0
+0 1
+
+
+//----------------- Allows for lagged values of the state variable to be encoded -----------------
+//== Number of lags encoded for state_variable[1] ==//
+1
+
+
+//== Number of states for state_variable[2] ==//
+3
+
+//== Each column contains the parameters for a Dirichlet prior on the corresponding
+//== column of the transition matrix.  Each element must be positive.  For each column,
+//== the relative size of the prior elements determine the relative size of the elements
+//== of the transition matrix and overall larger sizes implies a tighter prior.
+//== Transition matrix prior for state_variable[2]. (n_states x n_states) ==//
+ 1.0  1.0  1.0
+ 1.0  1.0  1.0
+ 1.0  1.0  1.0
+
+//== Free Dirichet dimensions for state_variable[2]  ==//
+2 2 2
+
+//== The jth restriction matrix is n_states x free[j].  Each row of the restriction
+//== matrix has exactly one non-zero entry and the sum of each column must be one.
+//== Column restrictions for state_variable[2] ==//
+1 0
+0 0
+0 1
+
+0 0
+1 0
+0 1
+
+0 0.5
+0 0.5
+1 0
+
+
+//----------------- Allows for lagged values of the state variable to be encoded -----------------
+//== Number of lags encoded for state_variable[1] ==//
+0
+
+
+//----------------- User-defined intial values of the transition matrix Q_k, mainly for debugging purposes -----------------
+//Non-zero integer: use the values under Initial: Transition matrix; 0: use Waggoner's default value which is the prior mean.
+//== indxInitializeTransitionMatrix ==//
+1
+
+
+// 0.5  0.5
+// 0.5  0.5
+//Call ReadTransitionMatrices(f,(char*)NULL,"Initial: ",model) in the main function.
+//If indxInitializeTransitionMatrix=0 or does not exist, the following values are not used and
+//  Waggoner's function initializes the value from the prior mean.
+//== Initial: Transition matrix[1] ==//
+ 9.7331908462961059e-01  1.5790335972902768e-02
+ 2.6680915370389413e-02  9.8420966402709720e-01
+
+
+
+//--- Initial guess
+ 9.7331908462961059e-01  1.5790335972902768e-02
+ 2.6680915370389413e-02  9.8420966402709720e-01
+//--- Converged results
+ 9.7885175490925935e-001  1.6213439424811054e-002
+ 2.1148245090740647e-002  9.8378656057518898e-001
+
+
+
+
+//== Initial: Transition matrix[2] ==//
+ 3.8872218833300276e-01  0.0                     3.185699602083008e-002
+ 0.0                     9.8480151496509816e-01  3.185699602083008e-002
+ 6.112778116669972e-001  1.519848503490184e-002  9.3628600795833983e-01
+
+
+//--- Initial guess
+ 3.8872218833300276e-01  0.0                     3.185699602083008e-002
+ 0.0                     9.8480151496509816e-01  3.185699602083008e-002
+ 6.112778116669972e-001  1.519848503490184e-002  9.3628600795833983e-01
+//--- Converged results
+ 4.3952235591344235e-001  0.0000000000000000e+000  3.6997582827338338e-002
+ 0.0000000000000000e+000  9.7128777689293189e-001  3.6997582827338338e-002
+ 5.6047764408655765e-001  2.8712223107068113e-002  9.2600483434532332e-001
+
+
+
+
+
+
+
+
+
+
+
+
+//********************************** Notes ******************************************//
+//----------------- Not working any more but used to have the user-defined intial values of Q_k -----------------
+//== Initial transition matrix for state_variable[12]. (n_states x n_states)  ==//
+0.99 0.01
+0.01 0.99
+
+The following no longer used in the new switch.c:
+line 564 switch.c.
+// Initial transition matrix[1]  //
+0.8 0.1 0.1
+0.1 0.8 0.1
+0.1 0.1 0.8
+
+
+For the restrictions of the absorbing type:
+X  X  0
+X  X  0
+X  X  1
+
+we have:
+
+// Free parameters in state_variable[1] //
+3 3 1
+
+// Column restrictions for state_variable[1] //
+1 0 0
+0 1 0
+0 0 1
+
+1 0 0
+0 1 0
+0 0 1
+
+0
+0
+1
+//--------------------------------------------------------------------------//

tzDocumentation/readme_use_dwVARestimate.prn

+Zeta in DW's output is Xi^2 in SWZ's JE article.  The prior on Zeta is gamma(a=1,b=1).  <<>> Check with Dan to be sure (17 Jan 2010 with Kirstin)
+   Sigma(k) = inv( A0(K) Zeta(k) A0(k)' )
+
+In the init_*.dat, the following is always true (hard-coded when using create_init_file.c):
+//== Specification (0=default  1=Sims-Zha  2=Random Walk) ==//
+1
+
+But if you want to change this, you must go into this file to do it manually.
+
+
+
+3by2 (e.g., 3 states for variances and 2 states for coefficients)
+
+sv 1         sv 2
+3 states    2 states
+A1-A3       B1-B2
+
+Block A1
+B1 and B2
+
+Block A2
+B1 and B2
+
+Block A3
+B1 and B2
+==>
+Overall Sates:  Individual Sates
+1.                 A1 B1
+2.                 A1 B2
+3.                 A2 B1
+4.                 A2 B2
+5.                 A3 B1
+6.                 A3 B2
+
+
+WriteTransitionMatrices(f_out,(char*)NULL,header,model);
+Write_VAR_Parameters(f_out,(char*)NULL,header,model);
+
+
+create_init_file <
+TZMatlab file> <DWSpecification file> file_tag name.
+Example
+exectuable datainp2_a_case2p2_sz5v_lp5011_d8507.prn specification2p2_3by2.dat test
+All these do is to read the SZ prior and the data and the starting point for the estimation.
+
+In Specification*.dat:
+//== Controlling states variables for coefficients ==//
+
+//== Controlling states variables for variance ==//
+
+
+In init_tag.dat, we have
+
+Mean of the prior.
+//== Initial: Transition matrix[1] ==//
+
+//== Initial: Transition matrix[2] ==//
+
+
+
+
+If -ft is chosen, -fs, -fr, and -rh will be all ignored.
+For -fr option, supply a different file with the data from //== Initial: Transition matrix[1] ==// on.
+    Perhaps change to //== Restart: ?????==//
+-rh "Restart: " in the command line.
+  or if I'm lazy, I would do the following instead:
+-rh "Initial: " in the command line.
+
+Another example:
+
+executable -ft tag -MLE -ce 1.03-6 -ii 1.5
+        //-ii can be a double.
+
+
+
+
+
+
+Input Command line:
+
+Group 1 (must specify most inputs).
+   Attempt to set up model from command line.  Command line options are the following
+
+   -di <directory>
+      If this argument exists, then all input files are in specified directory.
+
+   -ft <filename tag>
+      If this argument exists, then the following is attempted:
+
+         specification file name:  est_final_<tag>.dat
+         init/restart file name:   est_final_<tag>.dat with header="Posterior mode: "
+
+         specification file name:  init_<tag>.dat
+         init/restart file name:   est_intermediate_<tag>.dat with header="Iteration %d: "
+
+         (not yet implemented)
+         specification file name:  init_<tag>.dat
+         init/restart file name:   est_csminwel_<tag>.dat
+
+         specification file name:  init_<tag>.dat
+         init/restart file name:   init_<tag>.dat with header="Initial: "
+
+      Failure to load both the specification and restart/init files causes the routine to exit.
+
+   -fs <filename>
+      If this argument exists, then the specification file name is <filename>.  The argument -ft
+      takes precedence over -fs.
+
+   -fr <filename>
+      If this argument exists, then the init/restart file name is <filename>.  Must be used in
+      conjunction with the argument -fs.  The default value is the filename associated with the
+      argument -fs.
+
+   -rh <header>
+      If this argument exists, then the header for the init/restart file is <header>.  Must be
+      used in conjuction with the arguments -fr or -fs.  The default value is "".
+
+   If no command line options are given, then attemps to use a default input file
+   with the name "default.ini".  Returns one valid pointer to a TStateModel upon
+   success and null upon failure.
+
+
+Group 2 (either output or controlling execution of csminwel) (has a lot of default values).
+   Attempt to set up model from command line.  Command line options are the following
+
+   -do <directory>
+      If this argument exists, then all output files are put in the specified directory.
+
+   -fo <filename tag>
+      If this argument exists, then the output files are
+
+         est_csminwel_<tag>.dat
+         est_intermediate_<tag>.dat
+         est_final_<tag>.dat
+
+      The default value is the filename tag associated with the argument -ft if it exists.  Otherwise
+      it is "default".
+
+   -MLE
+      Find the maximum likelihood estimate
+
+   -PM (default)
+      Find the posterior mode
+
+   -cb <floating point number> (default = 1.0e-3)
+      Beginning csminwel exit criterion
+
+   -ce <floating point number> (default = 1.03-6)
+      Ending csminwel exit criterion
+
+   -ci <floating point number> (default = 0.1)
+      csminwel exit criterion increment multiplier
+
+   -ib <integer> (default = 50)
+      Beginning csminwel maximum iteration value
+
+   -ii <floating point number> (default = 2)
+      csminwel maximum interation increment multiplier
+
+   If no command line options are given, then attemps to use a default input file
+   with the name "default.ini".  Returns one valid pointer to a TStateModel upon
+   success and null upon failure.

tzDocumentation/readme_use_dwVARmhm.prn

+Example:
+
+swzmnhstage1 -fi MHM_input.dat -ft sz5v_2vby2m
+
+Command line:
+
+   Attempt to set up model from command line.  Command line options are the
+   following
+
+   -di <directory>
+      If this argument exists, then all input files are in specified directory.
+
+   -do <directory>
+      If this argument exists, then all output files are in specified directory.
+
+   -ft <filename tag>
+      If this argument exists, then the following is attempted:
+
+         1) specification file name:  mhm_final_<tag>.dat
+            mhm arguments file name:  mhm_final_<tag>.dat
+
+         2) specification file name:  mhm_intermediate_<tag>.dat
+            mhm arguments file name:  mhm_intermediate_<tag>.dat
+
+         3) specification file name:  est_final_<tag>.dat
+            mhm arguments file name:  -fi <filename>
+
+   -fi <filename>
+      If this argument exists, then additional mhm arguments are read from the
+      input file with the given filename.
+
+   -fs <filename>
+      If this argument exists, then the specification file name is <filename>.
+      The argument -ft takes precedence over -fs.
+
+   -fp <filename>
+      If this argument exists, then the posterior is read from <filename>.  Must
+      be used in conjunction with the argument -fs.  The default value is the
+      filename associated with the argument -fs.
+
+   -ph <header>
+      If this argument exists, then the header for the posterior file is
+      <header>.  Must be used in conjuction with the arguments -fp or -fs.  The
+      default value is "Posterior mode: ".
+
+   If no command line options are given, then attemps to use a default input file
+   with the name "default.ini".  Returns one valid pointer to a TStateModel upon
+   success and null upon failure.
+

tzDocumentation/testjunk.prn

+Example:
+
+swzmnhstage1 -fi MHM_input.dat -ft sz5v_2vby2m
+
+Command line:
+
+   Attempt to set up model from command line.  Command line options are the
+   following
+
+   -di <directory>
+      If this argument exists, then all input files are in specified directory.
+
+   -do <directory>
+      If this argument exists, then all output files are in specified directory.
+
+   -ft <filename tag>
+      If this argument exists, then the following is attempted:
+
+         1) specification file name:  mhm_final_<tag>.dat
+            mhm arguments file name:  mhm_final_<tag>.dat
+
+         2) specification file name:  mhm_intermediate_<tag>.dat
+            mhm arguments file name:  mhm_intermediate_<tag>.dat
+
+         3) specification file name:  est_final_<tag>.dat
+            mhm arguments file name:  -fi <filename>
+
+   -fi <filename>
+      If this argument exists, then additional mhm arguments are read from the
+      input file with the given filename.
+
+   -fs <filename>
+      If this argument exists, then the specification file name is <filename>.
+      The argument -ft takes precedence over -fs.
+
+   -fp <filename>
+      If this argument exists, then the posterior is read from <filename>.  Must
+      be used in conjunction with the argument -fs.  The default value is the
+      filename associated with the argument -fs.
+
+   -ph <header>
+      If this argument exists, then the header for the posterior file is
+      <header>.  Must be used in conjuction with the arguments -fp or -fs.  The
+      default value is "Posterior mode: ".
+
+   If no command line options are given, then attemps to use a default input file
+   with the name "default.ini".  Returns one valid pointer to a TStateModel upon
+   success and null upon failure.
+