How to use e in stata?

[Lucas Long]

Hello every one.

How can i run this equation in stata?


1/(1+e^-x)


I can calculate x separately, but how to use the function e^-x i am not very sure.

Thanks in advance.

[Joro Kolev]

Do not know what you mean by "run this equation", the function is -exp()-

. dis exp(-7)
.00091188

[Lucas Long]

Do not know what you mean by "run this equation", the function is -exp()-

. dis exp(-7)
.00091188

i need to generate a variable equal: 1/(1+e^-x), and I already calculated x.

SO should I just write the code:

gen wanted= 1/(1+exp(-x))

Thanks

[daniel klein]

You might want invlogit(); you probably do not want the result in a variable ...

[Lucas Long]

You might want invlogit(); you probably do not want the result in a variable ...

Hi Daniel, could you explain why should I not use the result in a variable?


I need to generate a variable Z, where Z=1/(1+e^-x),

and X was also generate by a equation: X=a+b+c+d+e.

In this situation, how should I run this?

Thank you so very much.

[daniel klein]

You gave very little information before. If X is a variable, you probably want Z to be a variable, too. What I was implying is that you do not want to store a constant in a variable; there are exceptions for this, too.

[Joro Kolev]

Yes Lucas, just write -gen wanted= 1/(1+exp(-x))-.

Or use -invlogit- like Daniel suggested.

. sysuse auto, clear
(1978 Automobile Data)

. gen wanted= 1/(1+exp(-rep))
(5 missing values generated)

. gen wanted2 = invlogit(rep)
(5 missing values generated)

. summ wanted*

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
      wanted |         69     .952005    .0504659   .7310586   .9933072
     wanted2 |         69     .952005    .0504659   .7310586   .9933072

[Joro Kolev]

invlogit(x) = exp(x)/{1 + exp(x)} = exp(x)/{exp(x)*(1/exp(x) + 1)} = 1/{exp(-x) + 1}

[Lucas Long]

invlogit(x) = exp(x)/{1 + exp(x)} = exp(x)/{exp(x)*(1/exp(x) + 1)} = 1/{exp(-x) + 1}

Thanks Joro! You are very helpful.