User Tools

Site Tools


sdist

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

sdist [2010/06/11 01:54] (current)
jochen created
Line 1: Line 1:
 +====== sdist - statistical distribution functions ======
  
 +===== Motivation =====
 +For many tasks in statistical hypothesis testing, it is necessary to find the ordinate value where the cumulative distibution function reaches a particular value (inverse of the CDF). These functions are available in Matlab'​s stats toolbox, but to allow users without this toolbox to use NeuroElf, freely available code has been edited and put into one compound function, ''​sdist''​.
 +
 +===== Function reference ('help sdist'​) =====
 +<​file> ​ sdist  - statistics distributions
 + 
 +  FORMAT: ​      d = sdist(type, variate, df1 [, df2 [, pin]])
 + 
 +  Input fields:
 + 
 +        type        either of '​betacdf',​ '​betainv',​ '​betapdf',​
 +                    '​fcdf',​ '​finv',​ '​fpdf',​ '​tcdf',​ '​tinv',​ '​tpdf'​
 +        variate ​    ​variate (or p-value)
 +        df1, df2    d.f. (or other required parameter)
 +        pin         ​logical flag, inverse from 0.9 to 0.1 logic
 + 
 +  Output fields:
 + 
 +        d           ​determined value of requested distribution function
 +</​file>​
 +
 +===== Usage examples =====
 +  * find the value required to threshold a t-map with 25 degrees of freedom at p<0.005 (uncorr.):<​code matlab>​sdist('​tinv',​ 0.995, 25)</​code>​
 +  * find the value required to threshold an F-map with (3, 20) degrees of freedom at p<0.01 (uncorr.):<​code matlab>​sdist('​finv',​ 0.99, 3, 20)</​code>​
 +  * find out if a given t-statistic (3.314, df = 115) is enough to meet a specific alpha criterion (0.001):<​code matlab>​sdist('​tcdf',​ 3.314, 115) >= 0.999</​code>​
sdist.txt ยท Last modified: 2010/06/11 01:54 by jochen