Integrated Model for Asynchronous Learning and Predictive Analytics for Enhanced Learner Experience

Introduction

With the advancements in information technology, Online learning or e-learning is becoming more accessible and it is being embraced and accepted as an education method. The COVID-19 pandemic's effects have led to a significant increase in the usage of e-learning due to its great temporal and geographic flexibility, low knowledge acquisition threshold, and wealth of learning resources. A major technological advancement in online learning is the adoption of Artificial Intelligence (AI) (Anderson et al., 1985; Baker, 2016; Roll et al., 2018; Seo et al., 2020b; VanLehn, 2011), where AI capabilities are being utilized to provide personalized learning experiences to learners. It is also assisting in automating the Assessments where the manual processes on instructors’ part are being taken care of by AI systems. All these benefits that AI can deliver were not possible if the associated data was unavailable. As it is well known that online learning is delivered through a Learning Management System (LMS) which captures all the learner data including their grades, activities, and performance. This data from LMS becomes the most important input for AI systems since all the insights about learner performance levels and behavioral aspects will be based on LMS data.

Despite the availability of several features and a variety of additional functionality on the LMS, instructors find it difficult to gauge their students' learning progress, raising concerns about the efficacy of e-learning. Hence, it becomes important that the educational institutions empower their LMS with AI capabilities to gain deep insights about the learner's performance. Getting access to such deep insights is an upcoming area which is popularly known as learning analytics (LA). LA ensures that the learner data is collected, measured, analyzed, and used for reporting purposes. The objective of using learning analytics is to understand and optimize the learning environment. The field itself has grown to the next level in the recent past where so many new areas have now become part of learning analytics. Learning Analytics now comprises assessment, LMS, educational research and technology from the academic side, AI, data science, data visualization and statistics from the technical and analytics side.

The most relevant sort of learning analytics is predictive analytics, which aims to predict future events based on prior and present data (Gandomi & Haider 2015). As the name implies, predictive analytics seeks to forecast future occurrences, patterns, and trends under varied settings (Joseph & Johnson 2013). It uses several techniques, such as regression analysis, forecasting, pattern matching, predictive modeling, and multivariate statistics (Gandomi & Haider 2015; Waller & Fawcett 2013). The goal of prediction is to forecast students' and teachers' activities to generate data that can assist teachers in making decisions (Chatti et al., 2013).

To answer the question, “What will happen next?” predictive analytics is applied. What interventions and preventive measures, for example, might a teacher apply to lower the failure rate? Herodotou et al. (2019) provided evidence on how teachers can use predictive analytics to support active learning. Predictive analytics, according to a large body of research, can help teachers improve their teaching practice (Barmaki & Hughes, 2015; Prieto et al., 2016; Prieto et al., 2018; Suehiro et al., 2017) as well as identify groups of students who may require additional support to achieve desired learning outcomes (Goggins et al., 2016; Thomas, 2018).