Effective ML-based quality of life prediction approach for dependent people in guardianship entities

This paper proposes an effective approach for predicting quality of life (QoL) for depen-dent individuals in guardianship entities. In addition, it aims to improve the QoL of people with intellectual disabilities. The proposed QoL prediction approach employs machine learning (ML) techniques to model the relationship between eight aspects of QoL and the corresponding QoL index. It determines whether or not a person needs assistance based on the index value. The pro-posed approach determines the priority of care (PoC) value for each aspect of a person. Based on PoC, the deficit aspect is determined, followed by the type of assistance a person requires, based on the decision priorities. It also generates a support report with suggested actions to highlight the level in that aspect. In addition, we train multiple ML models to predict the standard score (SS), which represents the support value related to the eight aspects of QoL. Finally, we use SS values to train an ML model to predict the support intensity scale (SIS). On a dataset compiled from guardianship entities, the proposed approach is validated. The QoL index, SS, and SIS prediction models achieve average R2 values of 0.9897, 0.9998, and 0.9977 with a standard deviation of 0.0051, 0.0002, and 0.0007, respectively.(c) 2022 THE AUTHORS. Published by Elsevier BV on behalf of Faculty of Engineering, Alexandria University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).