Analiza porównawcza wybranych technik eksploracji danych do klasyfikacji danych medycznych z brakującymi obserwacjami
Comparative analysis of selected data mining approaches to the classification of medical data with missing values (covariates)

Summary/Abstract: In implementation projects, in particular when analyzing medical data, it is quite often necessary to deal with tackle decision problems with missing values of specific variables (covariates). The aim of this paper is to perform a comparative analysis of selected data mining approaches, particularly simple and combined classifiers (ensembles) to solve classification tasks with missing data. The research was conducted using data mining techniques implemented in STATISTICA Data Miner and WEKA Machine Learning environments. The source data was extracted from a hospital data base of lung cancer patients treated surgically at Wrocław Thoracic Surgery Centre in the period 2000-2011.

• Details
• Contents