Data Mining: Healthcare Application

John Walsh, Jagiellonian University


Contemporary models in healthcare that employ data mining techniques are considered an innovative method to use information technology in a more behavioral oriented science. Healthcare research has been employing data mining and knowledge management systems for over a decade; however, analysts report on mixed results. Several models, such as artificial neural network approaches and genetic algorithms, represent a promising approach to support healthcare decision making, while more traditional data mining techniques have not presented a significant improvement in healthcare studies. In this paper, we use a forecasting model to make predictions about the behavior of patients using their health data. Incorporating the findings of current health research, a two-stage data mining clustering approach to forecast short term behavior is exhibited. Our results show that the model was able to identify different patterns and classes of patients that behave differently. These results are later used to produce various patient-care treatments and policies. Furthermore, we emphasize the applicability of this model as a tool to promote health awareness and reduce health costs.