Diagnostic methods have been an important tool in regression analysis to detect anomalies, such as departures from error assumptions and the presence of outliers and influential observations with the fitted models. Assuming censored data, we considered a classic analysis and Bayesian analysis assuming no informative priors for the parameters of the model with cured fraction. A Bayesian approach is considered using Markov Chain Monte Carlo Methods with Metropolis-Hasting algorithms steps to obtain the posterior summaries of interest. Some influence methods, such as the local influence, total local influence of an individual, local influence on predictions and generalized leverage are derived, analyzed and discussed in survival data with a cured fraction and covariates. The relevance of the approach is illustrated with a real data set, where it is shown that, by removing the most influential observations the decision about which model best fits the data is changed.
© 2008-2024 Fundación Dialnet · Todos los derechos reservados