public class ProportionalHazardsEx1 extends Object
Performs proportional-hazards data analysis on lung cancer data.
The following example is taken from Lawless (1982, page 287) and involves the survival of lung cancer patients based upon their initial tumor types and treatment type. In the example, the likelihood is maximized with no strata present in the data. This corresponds to Example 7.2.3 in Lawless (1982, page 367). The model is given as:$$\mathrm{ln}(\lambda)=\beta_1x_1+\beta_2x_2+\beta_3x_3+\alpha_i+\gamma_j$$
where \(\alpha_i\) and \(\gamma_j\)
correspond to dummy variables generated from classification variables in
columns 5 and 6 of x
. Respectively, \(x_1\)
corresponds to column index 2,
\(x_2\) corresponds to column index 3, and
\(x_3\)
corresponds to column index 4 of x
. Column 0 of x
contains the response and column 1 of x
contains the censoring
code. Logging is used to print output statistics.
Constructor and Description |
---|
ProportionalHazardsEx1() |
Copyright © 2020 Rogue Wave Software. All rights reserved.