Predicting the Flow of Patients in the Dental Industry

Abstract: In this paper, ML models are applied to discover the predictability of show-ups among patients visiting a specific company, Distriktstandvården. The motivation is to enable overbooking of scheduled time-blocks to increase the efficiency of the company’s use of resources. To achieve such efficiency, patient predictability has to be investigated and used. The application of the company’s internal data to train and test a machine learning model is presented and assessed in this paper. The result was that despite there being a large quantity of contributing parameters as well as a vast database to train on, the quality of the parameters and complexity of the task proved to only contribute to a precision quality of 69.3%, where 75% was deemed a desirable target. Despite the efforts in optimizing the model, and with the testing of alternative model candidates, the models did not discover strong links using the provided data for predicting patient show-up.