The Finitely Geometric Symptom Analysis in the Glioma Survival Study

Novel approach for statistical inference in survival analysis based on factor or dichotomic variables is proposed. We are seeking for the most informative finitely linear combinations (symptoms) of variables in the finite field. This procedure necessarily yields the new variable of the same nature (e.g. factor). Different measures can be used as an optimality criterion of such combination: entropy, the uncertainty coefficient, p-level of some statistical tests. We use this method to determine the major factors of the glioma postoperation survival. These factors were found to be age, high-grading factor, stage of illness and a factor of the stereotactic cryodestruction.