"Program" ml model - ml check error message: returned error 198

[Insa Flachsbarth]

Dear all,

I programmed the following ml model d0 which I somehow don't get to run, supposedly because of some invalid names. I am using a balanced panel dataset T=7, N = 187. Year is the variable defining the time dimension and countryid is the panel variable. Let me give you the code:

global MY_panel countryid
    program ziqcmle
    args todo b lnf
    tempvar xb zb last mu_it mu_i_ mu_i p_it y_i_ y_i sumqcl_i
    mleval ‘`xb' =‘ `b'’, eq(1)
    mleval ‘`zb’' =‘ `b'’, eq(2)
    local y_it "$ML_y1"
    local i $MY_panel
    * calculate the log-likelihood
    quietly {
    sort `‘i'’
    by ‘`i'’: gen double `‘last'’ = (_n==_N)
    gen double ‘`mu_it'’ = (exp(`‘xb'’)/(1+exp(`‘zb'’)))
    by ‘`i'’: gen double `‘mu_i_' = sum(`‘mu_it'’)
    by ‘`i'’: gen double ‘`mu_i’' = ‘`mu_i_'’[_N]
    gen double `‘p_it'’ = ‘`mu_it'’/`‘mu_i'’
    by ‘`i'’: gen double ‘`y_i_'’ = sum(`‘y_it'’)
    by ‘`i’': gen double ‘`y_i'’ = `‘y_i_'’[_N]
    by ‘`i’': gen double `‘sumqcl_i’' = sum(‘`y_it'’*ln(`‘p_it’') - ‘`p_it'’*`‘y_i’')
    mlsum `‘lnf'’ = ‘`sumqcl_i'’ if `‘last'’==1
    if (`‘todo'’ == 0 | `‘lnf'’==.) exit
    }
    end
    
    * Estimate the ZI-QCML model
    ml model d0 ziqcmle (y = x , nocons) (y = z)
    ml check        
    ml search    
    ml maximize

I now get the following error message after ml check (irrespective of which variables I use for y, x and z). So there seems to be some error with the line where

mleval ‘`xb' =‘ `b'’, eq(1)

.

Output after ml check:

ml check

Test 1: Calling ziqcmle to check if it computes log likelihood and
does not alter coefficient vector...
FAILED; ziqcmle returned error 198.

Here is a trace of its execution:

---

-> ziqcmle 0 __000009 __00000A __00000B __00000C
- `begin'
= capture noisily version 14.2: ziqcmle 0 __000009 __00000A __00000B __00000C
--------------------------------------------------------------------------------------------------------------------------- begin ziqcmle ---
- args todo b lnf
- tempvar xb zb last mu_it mu_i_ mu_i p_it y_i_ y_i sumqcl_i
- mleval ‘`xb' =‘ `b'’, eq(1)
= mleval ‘__00000D =‘ __000009’, eq(1)
‘__00000D invalid name
----------------------------------------------------------------------------------------------------------------------------- end ziqcmle ---
- `end'
= set trace off

---

Fix ziqcmle.
r(198);

end of do-file

I am not quite sure what I am doing wrong here, though I have to admit I am just starting with ml model programming. Does anyone have a hint?

Many thanks and please let me know if you need more information.

Best, Insa

[Insa Flachsbarth]

The problem is solved. Just for your information: I had copy pasted part of the code from another source. Somehow Stata had difficulties identifying the copied text. So I rewrote the code again and it worked. It is a bit strange as it looks exactly the same, but it solved the problem...
Just in case someone faces a similar issue at some point...