contents next up previous
Next: Non-Parametric Cdf, categorical variable Up: Basic classes Previous: Gaussian Cdf

Subsections

Non-Parametric Cdf, continuous variable

Non_param_cdf<lower_tail_interpol,middle_interpol,upper_tail,T>

A non-parametric cdf of variable Z is a cdf F defined by a discrete set of points $ \Big($zi, F(zi)$ \Big)$, i = 1,..., n, z1$ \le$...$ \le$zn. As Z is a continuous variable, the points $ \Big($zi, F(zi)$ \Big)$ must be interpolated in order to associate a probability to any z-value different from the zi's. Call z an outcome of Z different from zi (for all i). If z < z1 or z > zn, a function of type lower_tail_interpol or upper_tail_interpol is used to interpolate the cdf and compute F(z). If z1$ \le$z$ \le$zn a function of type middle_interpol is used. Similarly, when computing the inverse of the cdf for a probability p, if p < p1 or p > pn, a function of type lower_tail_interpol or upper_tail_interpol is used to interpolate the cdf and compute F-1(p). If p1$ \le$p$ \le$pn a function of type middle_interpol is used.



Where Defined

In header file <cdf.h>



Template Parameters

lower_tail_interpol   the type of the function used to interpolate the lower tail of the distribution
upper_tail_interpol   the type of the function used to interpolate the upper tail of the distribution
middle_interpol   the type of the function used to interpolate between two known values zj and zj + 1
T   the cdf's type value. It is double by default.



Model of

Non-Parametric Cdf



Type Requirements



Members


contents next up previous
Next: Non-Parametric Cdf, categorical variable Up: Basic classes Previous: Gaussian Cdf
nicolas
2002-05-07