Computation of the MLE for Bivariate Interval Censored Data
Data from the Aids Clinical Trials Group protocol ACTG 181
Modified data from the Aids Clinical Trials Group protocol ACTG 181
Transform (intersections of) canonical rectangles back to their origin...
Compute the MLE for bivariate censored data
Breast cosmesis data
Example data set (artificial)
Menopause data
Modified menopause data
Create a marginal CDF (or survival function) plot of the MLE
Create a bivariate CDF (or survival function) plot of the MLE
Plot a clique matrix
Create a univariate density plot of the MLE
Create a bivariate density plot of the MLE
Plot a height map
Plot a set of rectangles
Transform a set of rectangles into canonical rectangles
Determine areas of possible mass support of the MLE
We provide functions to compute the nonparametric maximum likelihood estimator (MLE) for the bivariate distribution of (X,Y), when realizations of (X,Y) cannot be observed directly. To be more precise, we consider the situation where we observe a set of rectangles in R^2 that are known to contain the unobservable realizations of (X,Y). We compute the MLE based on such a set of rectangles. The methods can also be used for univariate censored data (see data set 'cosmesis'), and for censored data with competing risks (see data set 'menopause'). We also provide functions to visualize the observed data and the MLE.