Applied Mathematics & Information Sciences

Author Country (or Countries)



n this paper, we illustrate the applicability of a regression model referring to patients with cutaneous melanoma. The reason for the choice of the dataset is due the disease be responsible for most cases of skin cancer. The proposed regression model is based on the Weibull-Poisson distribution in a structure of long-term modeling, in which the covariates were included in the proportion p of cured via the logistic link function. The motivation to study this distribution is since, in addition to generalizing the Weibull model, which is widely used in several areas of knowledge among them the survival analysis, it presents great flexibility in the forms of the hazard function. Through the proposed model we observed that the estimated proportion of cured is higher for patients with few nodules. In addition, it was presented the estimated hazard function of the patients and was verified that has a unimodal form.

Digital Object Identifier (DOI)