Pruebas no param´ etricas para comparar curvas de supervivencia de dos grupos que experimentan eventos recurrentes

The objective of this paper is to propose statistical to comp are survival curves of groups what experienced recurrent events. The idea comes from the comparison of the models of classical analysis and they are generalizations of the statistical comparison of the weighted traditional survival analysis. An event is recurring if it can happen more than once in the unit under study. Survival analysis techniques have traditionally had focused its study of a single occurrence of the event per unit. However, recent research has proposed several models to study such phenomena. These events are phenomena that occur in many areas. Infinities such events tend to occur in ou r environment include: viral diseases, cancer tumors, fevers, machinery and equipment failures, births, murders, earthquakes, rain, volcanic eruptions, industrial accidents, ca r accidents, among others. Recently, have been developing new techniques and models to study recurrent phenomena. In this research, are listed some of them and our aim is to compare survival curves between groups who experience these phenomena. We intend to generate statistical comparison to measure the st atistical differences in survival curves what have estimated with the nonparametric models. The idea consist in generalize classical tests weighted of the statistical analysis and will apply on recurrent events. In the investigation, we will design prog rams in R language to evaluate the proposed. In this work we will use the database the experiment of Byar, that is an experiment that measures the time (months) of recurrence of tumors of one hundred sixteen (116) patients with superficial bladder can cer treated with placebo, thiotepa and pyridoxine. The aim of this study consist in the application of the proposed to compare the survival curves of the three groups, and determine whether there are significant di fferences between treatments.