이산화 기법을 이용한 심장병 수술의 예후예측

Predicting Prognosis of Cardiac Surgery using Discrete Methods

In particular, use of big data in the medical field improves the quality of services provided to the patient and reduce costs to the hospital. Moreover, it is useful in predicting the effectiveness of the various surgical procedures. Data mining can extract information on the basis of big data and can be used to exlore mass data efficiently. In data mining, there are some special technology and classification technique. Therefore, for efficient data mining, it is important to select an efficient algorithm from various special techniques and classification techniques. However, to select the appropriate classifiers and algorithms is a difficult problem. In this paper, I tried to experiment with special technology that can improve the classification performance through the Cardiac Surgery by Surgeon Beginning2008 of US Government's open data that has been collected in mortality after heart surgery. I compared accuracy of supervised discretization with accuracy of unsupervised discretization using the nearest neighborhood classifiers. As a tool for experiment, I used WEKA v3.6.1 developed by Waikato University.