Recommending Personalized Interventions to Increase Employability of Disabled Jobseekers

Ha Xuan Tran, Thuc Duy Le, Jiuyong Li, Lin Liu, Jixue Liu, Yanchang Zhao, Tony Waters

Abstract: An emerging problem in Disability Employment Services (DES) is recommending to disabled jobseekers the right skill to upgrade and the right upgrade level to achieve maximum increase in their employment potential. This problem involves causal reasoning to estimate the causal effect on employment status to determine the most effective personalized intervention. In this paper, we propose a causal graph based method to solve the intervention recommendation problem. Personalized causal graphs of individual training samples are reverse engineered from a population-level causal graph using linear interpolation. A prediction model is built from these personalized graphs to recommend interventions. Experiments with a case study from an Australian DES provider show that by adopting interventions recommended by our method, disabled jobseekers would increase their employability by up to 24%. Evaluations with public datasets also show its advantages in other applications.