Year of Graduation
Knowledge discovery for the problem of comparing curing strategies
School of Applied Mathematics and Information Science
Nowadays leukaemia is not a fatal disease any more. Modern protocols for children acute lymphoblastic leukaemia allow physicians to reach a high ratio of fully recovered patients. However, it has become difficult to improve treatment for the whole set of patients. One of the most promising ways for therapy optimization is subgroup analysis. The problem is to find subgroups in which results of treatment strategies are significantly different. In this work an original approach to the study of different types of treatments in subgroups is proposed. This approach is based on matching theory and decision trees. It was developed for the data on children with acute lymphoblastic leukaemia, but can be easily adapted to other data. The observed results of the application to the initial data are presented.