An Integrated Survival Analysis Method Based on Location Scale Model

For survival analysis, the samples are always tested under various conditions, and the test results are extrapolated from the test conditions to the normal conditions. To assess the survivability as precisely as possible, a method of integrated survival analysis based on location scale model was proposed, using the test data under varied conditions. In this method, the location parameter was assumed as a function of relevant condition variables while the scale parameter was assumed as an unknown positive constant. Then the location parameter function was constructed by the method of radial basis function. Thus the influences of condition on the survivability were measured quantificationally, especially the interaction of varied condition variables. To draw a better inference, the method of shrinkage estimation was introduced to combine the sample information with relevant prior information to present the shrinkage estimation of location parameter. The instance analysis shows that this integrated survival analysis method is feasible and straightforward for engineering application.