Measuring the Impact of Automation on Occupational Skills in UAE's Labor Force Using Topic Models

Abstract

In recent years, the rapid advances of digital technologies and automation lead to profound changes in skill composition in the workplace. Middle and low skilled jobs are constantly being replaced by automated systems. Tasks requiring manual and perception skills decreased in the past decade while interpersonal and facility with technology increased. A recent study argues that susceptibility of jobs to automation should not be defined as a priori which could lead to a limitation by the assumptions inherent in logical inference. In this study, skill-dimensions are determined using factorization and clustering methods. The skill-dimensions are then evaluated using a pre-trained NLP model. To determine the magnitude of change in the workforce in the UAE, demand of jobs in the UAE labor have recently been developed in a data mining approach to collect a corpus of job advertisements which were mapped to O*NET, the most comprehensive publicly accessible database of occupational requirements. This study builds on the pre-generated UAE job demands to identify the changes in occupational skills by utilizing four well-established factorization and clustering techniques: Principal Component Factor Analysis (PCFA), Non-Negative Matrix Factorization (NMF), K-Means Clustering, and Density-based spatial clustering of applications with noise (DBSCAN). A sensitivity analysis to find the ideal parameters in eight simulations is developed to capture the skill-dimensions. Followed to that, three KPIs based on the pre-trained NLP model were developed to evaluate the quality of the computed dimensions. Jobs identified as important to the UAE using the pregenerated weights were then used to create skill-dimensions focusing on the following two industries: Banking and Finance, and Oil and Gas. The new industry-focused dimensions helped determine the impact of advances of technology on UAE workforce in the past 10 years. The study is concluded with implications to determine the jobs and skills which will have the most impact of advances in digital technologies and recommendations to ensure that UAE workforce is well equipped to move towards an independent knowledge-based economy that is equipped with the right knowledge and skill-sets to move towards emerging industries driven by automation.

Date of Award	Jul 2018
Original language	American English
Supervisor	Mohammad Omar (Supervisor)