Dependency, defined as the need for assistance in daily activities due to physical or cognitive limitations, is a crucial concern in healthcare and social services. Accurately appraising dependency levels is essential for allocating resources, planning care, and enhancing the quality of life for individuals needing support. This work details developing a system to assess and predict individuals’ dependency levels through an optimized survey. Initially, a comprehensive scale with 26 items was created by social work experts and tested with 210 users with three different dependency levels recognized and 40 without any dependency. A machine learning technique has been used to find a reduced set of items that predict the dependency level accurately. The results classify the fourth class with 91 % accuracy using only nine items. This system has been deployed and tested in a web application, so other systems or persons can use it to predict the individuals’ dependency levels online.
Buscar en Dimension