PREDICTION OF DISEASE STAGE IN PATIENTS WITH ADENOMYOSIS USING CLASSIFICATION TREE

Aim. This study was designed to develop the mathematical prediction model of adenomyosis spread stages according to the results of clinical examination using the classification tree statistical method.

Materials and methods. During this study we conducted the sampling of 84 patients with adenomyosis. By means of nonparametric correlation analysis we identified the indicators which were interconnected with the disease stage and prediction according to the results of clinical examination of the patients by means of the classification tree statistical method.

Results. We managed to build a suitable classification tree that helped to reach the compromise between the tree complexity and the amount of false classifications. This method allows us to define to role (significance) of the predictors in the classification model.

Conclusion. The creation of software applications automatizes the classification procedure and makes it possible for medical staff who don’t have specialized training in data analysis sphere to use it.