def hypergeoP_byLog(n,i,m,N):
"""
Calculates the non-cumulative hypergeometric p-value for variables:
n = # of positives in population
i = # of positives in sample
m = # of negatives in population
N = sample size
P(x=i) = (choose(n,i)choose(m,N-i))/choose(n+m,N)
For more details -> http://mathworld.wolfram.com/HypergeometricDistribution.html
"""
return 10**(log10(bestChoose(n,i))+log10(bestChoose(m,N-i))-log10(bestChoose(n+m,N)))
def hypergeoP(n,i,m,N):
"""
| Calculates the non-cumulative hypergeometric *p-value* for variables:
|
| ``n`` = number of positives in population
| ``i`` = number of positives in sample
| ``m`` = number of negatives in population
| ``N`` = sample size
|
| *P(x=i) = (choose(n,i)choose(m,N-i))/choose(n+m,N)*
|
| For more details -> http://mathworld.wolfram.com/HypergeometricDistribution.html
"""
return (bestChoose(n,i)*bestChoose(m,N-i))/float(bestChoose(n+m,N))
def binomialPval(n,k,p):
"""Returns exact binomial P-value.
n = number of trials
k = number of successes
p = probability of a success
P(k succeses in n trials) = choose(n,k) * p^k * (1-p)^(n-k)
"""
return bestChoose(n,k) * p**k * (1-p)**(n-k)
def binomialPval(n,k,p):
"""
*RETURNS:*
* exact binomial P-value.
| ``n`` = number of trials
| ``k`` = number of successes
| ``p`` = probability of a success
|
| *P(k succeses in n trials) = choose(n,k) (p^k) ((1-p)^(n-k))*
"""
return bestChoose(n,k) * p**k * (1-p)**(n-k)
