Neural-Network-Based Outcome Classification for Nursing Care

Abstract

As the number of aging populations grows rapidly, the demands of nursing care increase sharply. How to guarantee the nursing care quality becomes a great issue. Traditional formulation of nursing care plans relies on numerous experiences and professional knowledges. However, there are not enough qualified nursing workers. The development of nursing informatization and artificial intelligence provides a feasible solution. In this paper, we aim to build the machine learning model to predict the outcome domain only based on outcome indictors, so that when the nurses obtain the indictors after the evaluation, the model could output the predicted domain to guide the nursing care plan formulation. We extracted outcome domains and the corresponding outcome indictors from the guidelines. Then we applied the TF-IDF method to transform the indictors into numerical vectors, which were further used to train an artificial neural network (ANN) model. In the experiments, 481 outcomes were extracted, which belonged to 6 domains. Besides, to validate the model, we compared it with KNN, Support Vector Machine, and Random Forest. 10-folds results showed that ANN achieved the best accuracy, i.e. 84%, which proves the feasibility of predicting the outcome classification only based on the indicators, and using machine learning to help make nursing plans.