Design and development of healthcare systems for outpatients using big data and predictive analytics

Abstract

Healthcare service for outpatients is a patient-oriented service, and is gradually becoming an important aspect in healthcare systems. Recommender Systems and Appointment Scheduling Systems are important facets in the healthcare systems. Healthcare Recommender Systems are a class of information filtering systems that help users to discover items that might be of interest to them. In this thesis, two main entities in the MPRS-PS healthcare systems are patients and physicians. The MPRS (Multi-criteria Physician Recommender System) needs to assist patients in making decisions regarding the physicians best suited for the patient s need. The PS (Patient Scheduler) schedules patients in such a way that the patient s priorities and patient s preferences are taken into consideration. The thesis presents a detailed view of both Recommendation Systems and Appointment Scheduling systems describing the limitations of current methods in a big data healthcare environment. newlineA Multi-criteria Physician Recommender system is a technique used to predict unknown ratings and recommend items to patients based on ratings given for the doctors. The multi-criteria system combines machine learning and deep learning techniques for modelling preferences of patients (users) based on several attributes of the physicians (items). The proposed MPRS system is based on machine learning techniques to predict individual ratings and deep learning techniques for predicting the overall ratings. In the experiments, the performance of the MPRS model is evaluated in a hospital and the efficiency evaluated. newlineAppointment Scheduling is a wide area of research and various factors need to be considered while scheduling a patient to a physician. However, in traditional healthcare systems, all patients are considered to be in the same priority, which results in patients not showing up for the appointment and resources are unused. newline