An exploratory study in classification methods for patients´ dataset

There are various methods in data mining that can be applied in classification data. This paper discusses the experiments done in classifying ICU data. The dataset consists of 25 variables for 410 patients. The goal of this experiment is to determine the survival of the patients, so the targeted output are alive and dead. Three selected data mining methods are decision tree, Naives Bayes and logistics regression. Based on mean absolute error and root-squared error, the later method provides a better result. The result of classification could be used to help hospitals in predicting their patients´ status and provide better way of antibiotic treatment. Applying an intelligent tool to classify the antibiotic resistance may support the decision making to diagnose the patients in an effective way. A right treatment will make sure the patient is survived. This intelligent tool for managing medicine dosage is worthy and brings a huge impact to medical sector.