AScheme to Enhance the Employability of the Differently-Abled

World Health Organisation reports indicate that 15% of the world's population is differently - abled. Employing differently - abled individuals will not only be a service to the society, but also help to improve the Gross Domestic Product (GDP) of the country. The type of jobs a differently - abled person can perform depends on factors like the type of disability as well as the degree of disability, the education and skills of the person. This paper proposes a scheme which recommends the type of training for the differently - abled to enhance his or her employability. The proposed scheme uses reinforcement learning to develop a model that recommends a viable course that can be taken up by a person with a disability, given, his or her profile. Although there is research on using reinforcement learning for course recommendation, as per our understanding, no attempt has been made to develop a course recommendation system for the differently-abled using reinforcement learning. The results of this research is presented using the following graphs, namely episodic reward mean, explained variance, entropy loss, clip fraction and approximate Kullback-Leibler (KL) divergence. These graphs establish that the model has learned effectively and has reached a stable policy, indicating that the value function is increasingly accurate but is not overfitting.