python - Fisher's exact test for bigger than 2 by 2 contingency table -
hi scipy stats has implementation of fisher's exact test 2 2 contingency tables. want test on bigger 2 2 tables. (5x2 ,5x3) know there fisher.test in r can job want in python code
anybody knows python implementation of fisher's exact test can work on bigger tables?
also not sure if ok fisher's exact test on bigger 2 2 tables.
thanks
yes, ok fisher's exact test on 5x2 or 5x3 tables.
there aren't clean, tested solutions out there in python. 1 solution use rpy2 , call r function python.
import rpy2.robjects robjects rpy2.robjects.packages import importr stats = importr('stats') v = robjects.intvector([4,4,10,4,5,6]) m = robjects.r['matrix'](v,nrow=3) res = stats.fisher_test(m) print m >>> [,1] [,2] >>> [1,] 4 4 >>> [2,] 4 5 >>> [3,] 10 6 print 'p-value: {}'.format(res[0][0]) >>> p-value: 0.668165917041
another solution dig c code r version uses , call code directly. here link someone's github project went original fortran implementation , call python.
Comments
Post a Comment