Student Outcomes Assessments Using Deep Learning

Abstract

The pursuit of academic accreditations for degree programs is a common objective among universities worldwide. This objective is driven by the recognition that accreditation not only enhances the quality of teaching within the institution but also facilitates the recruitment of highly qualified faculty members and students. In fact, accreditation has become a near-universal requirement for universities across the globe. Even ABET, the accrediting body primarily responsible for institutions in the United States, has expanded its scope to include programs on a global scale. A significant portion of the documentation submitted to accreditation agencies pertains to the collection and reporting of data on student achievement of course learning outcomes (CLOs), Program Learning Outcomes (PLOs), and Key Performance Indicators (KPIs). Given the advancements in big data and the recent progress in data mining and machine learning, it is imperative to establish a methodology for data reporting and an automated evaluation system to effectively measure performance indicators. This research paper proposes the utilization of an intelligent system based on the Artificial Neural Networks (ANN) model to assess the ABET-defined Student Outcomes (SO) through the classification of their associated Key Performance Indicators (KPIs) at the program level. The proposed model employs deep learning techniques with the Multilayers Perceptron classifier (MLP) and comprises four layers: the input layer, two hidden layers, and the output layer. The findings presented in this paper serve as a proof of concept for the feasibility of an intelligent system that can generate meaningful data relevant to the accreditation process, regardless of the size of the academic department.