Identifying Reliable Predictors of Educational Outcomes Through Machine-Learning Predictive Modeling

Abstract

Results-based financing has guided the development of policies with measurable results improving learning outcomes at micro/macro levels. However, it is then necessary to identify factors which predict early and accurately favorable or challenging conditions for learning. Learning outcomes depend on complex interactions between multiple variables, many of which are not fully understood. The objective was to develop valid and accurate models predicting low and high levels of math performance and Vietnamese language, using machine-learning algorithms, as part of an international large-scale project in primary education in Vietnam. The models achieved very high accuracy (95-100%). A strong common pattern has been found for both Math and Vietnamese language, for the low and high levels of performance: the individual cognitive characteristics, physical factors and daily routines/ activities of the child are very important predictive factors of academic performance, as measured by student performance in the final Grade 5 test in math and Vietnamese, respectively. Parental expectations, preschool attendance and school trajectory of students have added relative importance in the classification. In order to accurately identify an expected low or high academic performance outcome, it is the full pattern of variables contained in the vector of information from each case that should be considered. Because, although each variable in a particular vector has a small contribution to the total predictive weight, it is the overall pattern containing the interactions between these variables that carries the necessary information for the accurate predictions. In addition, the identification of specific patterns for extreme groups of performance provides the necessary guidance for more focused educational interventions/investment and sound educational policies.

Modeling the Contribution of Genetic Variation to Cognitive Gains Following Training with a Machine Learning Approach

“…Some authors have shown that traditional statistical methods do not always yield accurate predictions and/or classifications (Bansal, Kauffman, & Weitz, 1993; Duliba, 1991; Everson, 1995). A more robust and accurate approach has been developed during the last decade and applied in health and education fields for the purpose of prediction (Cascallar, Boekaerts, & Costigan, 2006; Everson, Chance, & Lykins, 1994; Gorr, 1994; Hardgrave, Wilson, & Walstrom, 1994; Musso & Cascallar, 2009; Musso et al, 2012; Musso, Kyndt, Cascallar, & Dochy, 2013; Musso, Cascallar, Bostani, & Crawford, 2020; Musso, Hernández, & Cascallar, 2020). Machine‐learning techniques, such as methods using artificial neural networks (ANN), have been shown to be very effective to study problems consisting of a large number of variables in complex, nonlinear, and poorly understood interactions (Cascallar, Musso, Kyndt, & Dochy, 2014).…”

A prediction model for benefitting e-commerce through usage of regional data: A new framework

“…Singhal et al [36] proposed a logical structure to improve presentation of a web page, making it more userfriendly to promote e-business. Musso et al [37] used machine learning predictive models to predict low and high levels of performances in subjects in educational systems. Márquez-Chamorro et al [38] summarises about basic concepts of predictive monitoring area in business processes.…”

“…But unfortunately, more often these skilled persons are not that much educated overtime. Due to low level of literacy, they could not match themselves with the fast pace of online business applications and their practice [37]. Hence, they are lagging behind financially.…”

What predicts legislative success of early care and education policies?: Applications of machine learning and Natural Language Processing in a cross-state early childhood policy analysis

“…Topic Modeling [57] is a rapidly emerging area of research in educational data mining. Text analysis methods in education research have been used to 1) measure latent dispositions such as attitudes and beliefs of learners and instructors [60]; 2) explore the underlying topics and topic evolution spanning the 50-year history of educational leadership research literature [61]; 3) microclassroom processes such as MOOC interaction data [62]; 4) policy implementation and reform strategies [63]; or 5) identifying 'topics' of multiple setting-level predictors of student's language and math achievement outcomes [64]. Despite the promising potential of applying topic modeling in a variety of fields in the social sciences, its scalable, algorithmic approach to large-scale text data has received little attention among early childhood and education policy scholars [61].…”

“…However, a burgeoning body of work has applied machine learning algorithms in educational research to better understand how teaching and learning can be enhanced for whom under what conditions [60,89]. Recent studies, for example, have built artificial neural network models that predict students' language and math performance at large scale [64], introduced various methods in educational data science (EDS) for examining students' massive open online courses (MOOCs) interactions [62], or using LDA to analyze text data from thousands of school improvement reports to identify reform mechanisms that reduced student chronic absenteeism and improved achievement [63]. As such text mining can enhance administrative data that many early childhood researchers have been utilizing for decades as well as address barriers and limitations surrounding administrative data by adding richness that is often buried in digital text without the cost of primary data collection.…”