Predictive Analytics- A Target on Predicting the Loss of Patients in Healhcare Organizations
Healthcare is rich in information but poor in knowledge. This is because; there is no model to understand the behaviour of the patterns within the information. Data mining techniques can be leveraged to fill this gap. We use Regression tree, a powerful and popular tool for classification and prediction to predict the patients getting a heart disease. For this purpose we require a patient?s age, BP level, Cholesterol level and few more attributes. Health Risk assessment methods are basically about predicting the high-risk patients with their administrative claim data and medical diagnosis data. Effective intervention can predict the patients most at risk and reduce their cost of treatment. We develop a model that uses the previous year?s data as training variable and predict the LOS for a patient for the fore coming year. Accuracy of our model is measured by using the RMSE, difference between the estimate and actual value.
Keywords: Length Of Stay (LOS), Predictive Analytics, Framingham Risk Score, Regression Tree.
Conference Name: 11th International Conference on Science Engineering and Technology
Conference Date: 03, November 2015 - 04, November 2015
Paper ID: SET001