Hospital Length of Stay Prediction using Regression Models

The increasing healthcare costs and the availability of large volume of medical data motivate the search for ways to increase the care efficiency. The Length of stay[LOS] of hospitalized patients is considered as an important factor for healthcare policy planning as it has a major impact on the technical and financial resources as well as facilities occupation. It also indicates the hospital service quality. The objective of the proposed work is to build a Decision support system that can assist the physicians to predict the inpatient hospital LOS using the regression models namely Linear, Ridge, Lasso and ElasticNet. The comparative results of these regressors is presented with Mean Absolute error as the evaluation metrics. A GUI has been built as the application of the proposed work.