Ignore:
Timestamp:
Nov 12, 2008, 11:10:52 PM (15 years ago)
Author:
Peter
Message:

fixes #461. Also modified implementation of cdf_hypergeometric_P, which may cause conflict with modifications done in trunk (refs #87). If so, go with the trunk version (which uses GSL 1.8).

File:
1 edited

Legend:

Unmodified
Added
Removed
  • branches/0.4-stable/yat/statistics/Fisher.h

    r1621 r1624  
    118118    const unsigned int& minimum_size(void) const;
    119119
    120     ///
    121     /// If oddsratio is larger than unity, two-sided p-value is equal
    122     /// to 2*p_value_one_sided(). If oddsratio is smaller than unity
    123     /// two-sided p-value is equal to 2*(1-p_value_one_sided()). If
    124     /// oddsratio is unity two-sided p-value is equal to unity.
    125     ///
    126     /// If all elements in table is at least minimum_size(), a Chi2
    127     /// approximation is used.
    128     ///
    129     /// @return 2-sided p-value
    130     ///
     120    /**
     121       If all elements in table is at least minimum_size(), a Chi2
     122       approximation is used.
     123       
     124       Otherwise a two-sided p-value is calculated using the
     125       hypergeometric distribution
     126       \f$ \sum_k P(k) \f$ where summation runs over \a k such that
     127       \f$ |k-<a>| \ge |a-<a>| \f$.
     128
     129       \return two-sided p-value
     130    */
    131131    double p_value() const;
    132132   
Note: See TracChangeset for help on using the changeset viewer.