topic Re: SAS Datastep code example for probablility calculation please? in SAS Programming

SAS Datastep code example for probablility calculation please?

CharlotteCain — Mon, 10 Nov 2014 13:30:26 GMT

Hi Folks,

Can anybody provide me a simple SAS datastep algorithm for the below question and a statistical explanation for my understanding please

If I have a decision to make how do I calculate the probability of a selection how does one calculate the probability?

For example: What is the probability of me selecting Tea – with sugar – with milk.

What is the formula? 0.5 x 0.5 x 0.5

Re: SAS Datastep code example for probablility calculation please?

Ksharp — Mon, 10 Nov 2014 14:30:35 GMT

Sorry. I couldn't see that picture. Maybe it is being block by China.

Re: SAS Datastep code example for probablility calculation please?

PGStats — Mon, 10 Nov 2014 14:48:50 GMT

I can't see the picture either. I get the message

Cycle Prohibited

Description: Could not process your request for the document because it would cause an HTTP proxy cycle. Please check the URL and your browser's proxy settings.

Re: SAS Datastep code example for probablility calculation please?

CharlotteCain — Mon, 10 Nov 2014 15:36:15 GMT

Hi,

First off, I apologise for that coz it aint opening. Well the picture depicts as follows:

Tea ii. with sugar iii. With milk or iii. Without milk
Tea ii. without sugar iii. With milk or iii. Without milk
Coffee ii. With sugar iii. With milk or iii. Without milk
Coffee ii. Without sugar iii. With milk or iii. Without milk

Wish I could get the pic to work, but that's exactly the information.

Hope you are well.

Thanks so much,

Charlotte

Re: SAS Datastep code example for probablility calculation please?

stat_sas — Mon, 10 Nov 2014 17:52:48 GMT

Hi,

Probability selecting Tea – with sugar – with milk is

0.5 x 0.5 x 0.5

Which is aleardy mentioned above in your question.

Re: SAS Datastep code example for probablility calculation please?

CharlotteCain — Mon, 10 Nov 2014 18:32:45 GMT

Well, i noticed that. I'm not understanding the logic though or may be I'm having a blonde moment. lol. I thought it would be 0.5*0.25*0.25*0.25*0.25. I still would like you to explain me the formula if you don't mind.

Re: SAS Datastep code example for probablility calculation please?

stat_sas — Mon, 10 Nov 2014 18:41:49 GMT

Sure, tea – with sugar – with milk are 3 independent events each having two possible outcomes. In the given scenario with sugar has prob=0.5, with sugar prob=0.5 and with milk prob=0.5. In the next step we just need to multiply them.

Re: SAS Datastep code example for probablility calculation please?

Reeza — Mon, 10 Nov 2014 18:44:49 GMT

This is a better question for CrossValidated than here.

P(Tea, sugar, milk)= p(tea)*p(sugar)*p(milk)=0.5*0.5*0.5=1/8

Each odd is 0.5 because there are two options

i.e. choices are tea or coffee so p(tea)=1/2=0.5

and the same logic applies for remaining ingredients.

Another way to calculate this is to list all the possible combinations and see how many times you get your output:

1 out of 8 possible combinations:

Tea|Sugar|Milk

Tea|Sugar|No Milk

Tea|No Sugar|Milk

Tea|No Sugar|No Milk

Coffee|Sugar|Milk

Coffee|Sugar|No Milk

Coffee|No Sugar|Milk

Coffee|No Sugar|No Milk

Re: SAS Datastep code example for probablility calculation please?

ballardw — Mon, 10 Nov 2014 20:50:49 GMT

Pedantic note: Assumes the probabilities of the choices are independent.

Re: SAS Datastep code example for probablility calculation please?

Reeza — Mon, 10 Nov 2014 21:04:22 GMT

True

Re: SAS Datastep code example for probablility calculation please?

Haikuo — Mon, 10 Nov 2014 21:52:42 GMT

SAS has a function for it:

data _null_;

p=0.5; /*is a numeric probability of success parameter, the success rate of each test*/

n=3; /*is an integer number of independent Bernoulli trials parameter: the total number of the tests*/

m=3;

/*is an integer number of successes random variable: the number of successful test, in your case,

the m is the Maximum: n, as the probability of one success in one trial is the same as all success in all trials*/

prob=PROBBNML(p,n,m)- PROBBNML(p,n,m-1);

put prob=;

run;

Haikuo

Re: SAS Datastep code example for probablility calculation please?

Rick_SAS — Tue, 11 Nov 2014 01:21:01 GMT

Or use the CDF function. The following statements are equivalent:

prob=PROBBNML(p,n,m) - PROBBNML(p,n,m-1);

prob2 = cdf("Binomial",m,p,n) - cdf("Binomial",m-1,p,n);

Re: SAS Datastep code example for probablility calculation please?

Haikuo — Tue, 11 Nov 2014 01:37:23 GMT

Thanks for sharing, Rick! CDF function is obviously much more powerful.

Re: SAS Datastep code example for probablility calculation please?

CharlotteCain — Tue, 11 Nov 2014 04:54:54 GMT

Hi Haikuo and Rick,Thank you so much for the Genius help. I can't appreciate enough. Anyways, did you memorise SAS functions or you developed a knack to quickly search for the right now? I am asking this coz your inputs will only help me learn faster.

Cheers,

Charlotte

Re: SAS Datastep code example for probablility calculation please?

Rick_SAS — Tue, 11 Nov 2014 18:11:46 GMT

SAS is big. I don't pretend to know it all, but I've taken the time to learn about the functions that I use over and over. CDF is one of those functions. If you do a lot with univariate distributions in SAS, read this article: Four essential functions for statistical programmers - The DO Loop