New distribution families and their statistical properties for survival analysis

Survival analysis, which has an important place in many disciplines, involves the analysis issues of elapsed time period until a particular event occurs (failure time, or survival time). Therefore, properly describing the distribution of a 'survival time' as random variable is important to get the accurate estimates on survival and hazard rate. In this thesis, by giving the basic concepts of survival analysis, we introduce new generalized distributions, which can be used in survival and reliability analysis, based on the T-X method. We have studied the statistical properties of these distributions such as the density, hazard rate, survival, quantile, moment functions and maximum likelihood method for parameter estimation. Various real data applications are conducted to show the performance of the new distributions. It is concluded that according to goodness of fit tests, the newly introduced distributions outperform existing base distribution and alternative distributions in modelling survival data.