dest.py

from scipy import size,linspace
from matplotlib.pyplot import plot,show
from scipy.stats.kde import gaussian_kde

class dest(object):
    """Density estimator"""

    def __init__(self, rvs):
        self.rvs = rvs.copy()
        self.rvs.sort()
        self.size = size(self.rvs)
        self.kde = gaussian_kde(self.rvs)
    
    def pdf(self,x):
        if size(x)==1:
            return self.kde(x)[0]
        else:
            return self.kde(x)
        
    def ppf(self,x):
        if x<0 or x>1:
            raise ValueError("Not between 0 and 1")
        return self.rvs[int(x*(self.size-1))]
        
    def cdf(self,x):
        return size(self.rvs[self.rvs<x])/float(self.size)
        
    def graph(self,inc=1000):
        lowb = self.rvs[0] - (self.ppf(.1)-self.ppf(0.))
        upb = self.rvs[self.size-1] + (self.ppf(.1)-self.ppf(0.))
        x = linspace(lowb,upb,inc)
        plot(x,self.kde(x))
        show()
        
    def Confident(self, confLevel):
        lowerbound = int(round(((1.- confLevel)/2.)*self.size))
        upperbound = int(round((((1.- confLevel)/2.)+confLevel)*self.size))
        LowBound = self.rvs[lowerbound]
        UpBound = self.rvs[upperbound]
        return LowBound, UpBound