Advances and Applications in Statistics
Volume 64, Issue 1, Pages 1 - 22
(September 2020) http://dx.doi.org/10.17654/AS064010001 |
|
APPLICATION OF KUMARASWAMY INVERSE WEIBULL DISTRIBUTION IN MODELING SURVIVAL DATA
Dorcas W. Nduati, Levi N. Mbugua and Moses W. Wakoli
|
Abstract: The Weibull family of distributions has theoretical justification and is flexible in many applications because of its ability and shape in modeling a great selection of hazard rates. It also has an extended probability density distribution commonly used in modeling distribution of damaging potential of materials. Finally, it has recently been given priority because of its versatility to handle lifetime data. Minimal studies on survival time and hazard rates have utilized the Weibull distribution and other related distributions. Exploring mixing distributions that are beyond the Weibull distribution have not been widely utilized. In this study, Kumaraswamy inverse Weibull distribution was adopted to study survival information. Models with all risk factors were fitted for three distributions as Kumaraswamy inverse Weibull, log logistic and log normal distributions, respectively. The model with least Akaike information criterion, corrected Akaike information criterion and Bayesian information criterion was taken to be the best model. Kumaraswamy inverse Weibull distribution was a better distribution for fitting primary biliary cirrhosis data compared to log normal and log logistic distributions. Since Kumaraswamy inverse Weibull distribution is superior than Weibull distribution, the fit is the best for primary biliary cirrhosis data. Properties of Kumaraswamy inverse Weibull were got from method of moments. The final model had age and edema as risk factors but needed a medical doctor to advice on the biological causes of primary biliary cirrhosis. May be further research should be done and more variables could be added to the model to determine fully the cause of primary biliary cirrhosis. |
Keywords and phrases: inverse Weibull distribution (IW), Kumaraswamy inverse Weibull distribution (KIWD), method of moments, survival data, primary biliary cirrhosis (PBC).
|
|
Number of Downloads: 218 | Number of Views: 562 |
|