Survival Analysis of Lung and Bronchus Cancer Patients Segmented by Demographic Characteristics

摘要: We propose a Weibull mixture model considering both covariates and unobserved heterogeneity to examine how demographic variables affect individual survival times derive the annual number of deaths. analyze records patients diagnosed with lung bronchus cancer, most common cancer in United States. The result shows that diagnosis year as well age, gender, race, registry significantly times. including heterogeneity, we remove bias hazard rates provide better performance forecasting deaths than other benchmarks. Furthermore, from segmenting into several groups, specify difference between groups assess their group-specific probabilities within given period. Our study is distinctive bottom-up strategy adopted predict aggregate-level units. This makes health available two sides: public private sector. For sector, our enables more precise allocation government’s welfare budget. Also for segmentation results guidance insurance industry targeting customers specifically.