On Data Mining Technique for Differential Diagnostics Based on Data of Arterial Oscillography

Abstract

The work is devoted to development and application of the decision tree induction algorithm for the problem of differential diagnostics and assessment of adaptation possibilities of patients (cardiovascular, nervous, endocrine, respiratory system diseases) and healthy (in the position of lying and sitting) on the basis of the data of arterial oscillography. Software implementation uses C5.0 algorithm. For each type of research, as well as their totality, separate decision trees are constructed. The classification error as well as attribute usage are gotten and analyzed.