The Effect of Model Granularity on Student Performance Prediction Using Bayesian Networks

Pardos, Zachary A.; Heffernan, Neil T.; Anderson, Brigham; Heffernan, Cristina

doi:10.1007/978-3-540-73078-1_60

Cited by 27 publications

(18 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Thus, student performance prediction models based on Moodle log data have been proposed in multiple previous studies [5][6][7]. Additionally, log data from intelligent tutoring systems (ITS) have also been used for performance prediction [8]. In contrast, students' engagement with social media tools in emerging social learning environments has been less investigated as a potential performance predictor [9].…”

Section: Introductionmentioning

confidence: 99%

Predicting Academic Performance Based on Students’ Blog and Microblog Posts

Dascălu

Popescu

Becheru

et al. 2016

Adaptive and Adaptable Learning

View full text Add to dashboard Cite

Abstract. This study investigates the degree to which textual complexity indices applied on students' online contributions, corroborated with a longitudinal analysis performed on their weekly posts, predict academic performance. The source of student writing consists of blog and microblog posts, created in the context of a project-based learning scenario run on our eMUSE platform. Data is collected from six student cohorts, from six consecutive installments of the Web Applications Design course, comprising of 343 students. A significant model was obtained by relying on the textual complexity and longitudinal analysis indices, applied on the English contributions of 148 students that were actively involved in the undertaken projects.

show abstract

Section: Introductionmentioning

confidence: 99%

Predicting Academic Performance Based on Students’ Blog and Microblog Posts

Dascălu

Popescu

Becheru

et al. 2016

Adaptive and Adaptable Learning

View full text Add to dashboard Cite

show abstract

“…Bayesian networks in particular have been used to predict student applicant performance [12], to model user knowledge and predict student performance within a tutoring system [13] and also to predict a future graduate"s Cumulative Grade Point Average based on the students background at the time of admission [14]. Bayesian networks have also been used to model two different approaches to determine the probability a multi skill question has of being correct [15] and to predict future group performance in face-to-face collaborative learning [16].…”

Section: Overview Of Literaturementioning

confidence: 99%

Predicting Students’ Academic Performances – A Learning Analytics Approach using Multiple Linear Regression

Oyerinde¹,

Chia²

2017

IJCA

View full text Add to dashboard Cite

Learning Analytics is an area of Information Systems research that integrates data analytics and data mining techniques with the aim of enhancing knowledge management and learning delivery in education management..The current research proposes a framework to administer prediction of Students Academic Performance using Learning Analytics techniques. The research illustrates how this model is used effectively on secondary data collected from the Department of Computer Science, University of Jos, Nigeria.Multiple Linear Regression was used with the aid of the Statistical Package for Social Sciences (SPSS) analysis tool. Statistical Hypothesis testing was then used to validate the model with a 5% level of significance.

show abstract

Section: The Need For Flexible Support Of Competenciesmentioning

confidence: 99%

“…Such a classification needs to be reusable, as reproducible as possible, and coherent in order to enable reuse of LOs. In concrete domains it should reflect an optimal granularity to support instructional effectiveness [16] and it should be based on a cognitive task analysis. Now, our long-term experience shows that for pedagogists as well as for TELcontent it seems that different groups of designers and applicants are biased towards different classifications of competencies.…”

Section: The Need For Flexible Support Of Competenciesmentioning

confidence: 99%

“…A competency system should enable the system to infer knowledge/ability gaps concerning certain learning goals as well as subsequent steps in the actual learning/training session. In recent years, efforts in researching competencies have been multifold and often interdisciplinary, ranging from learning system/content specific skills [16] to approaches aiming at more abstract and general definitions of competencies [1,2,13]. Some researchers even include qualitative mappings between their competency system and other established CSs [15].…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Supporting Flexible Competency Frameworks

Melis

Faulhaber

Doost

et al. 2010

Advances in Web-Based Learning – ICWL 2010

View full text Add to dashboard Cite

Abstract. Since Bloom's initial work on competencies in 1956, various competency systems have been designed and used to assess students' competencies. Different pedagogical researchers and stakeholders prefer different systems. We have been collaborating with them. Such systems are essential for the adaptation by adaptive intelligent tutoring systems. Now, this paper presents how ActiveMath integrates several competency systems to bridge the gap between different competency systems and thereby facilitating the reuse of learning objects across system boundaries. The combination of competency-related data is achieved by mapping a new competency system to the internal one.Keywords: Competency System, Student Model, Course Generation The Need for Flexible Support of CompetenciesIn our long-term experience with metadata in eLearning and Intelligent Tutoring with ActiveMath [11], required competencies and difficulties of exercises turned out to be important not only for technical reasons but also in the interdisciplinary work with pedagogists.Pedagogists (and psychologists) have been looking for and using a classification system at least since 1956 [3]. The psychology-driven approaches in [5,1] and [2] are examples for later approaches most of which are empirically driven.The large PISA-studies needed a common classification for students' skills or competencies in order to compare student results across schools, regions, and countries. It seems that, similar to Bloom's taxonomy, teachers (and hence authors) could understand and apply the PISA competencies and even levels. 4 This seems to be more difficult for the more complex taxonomy in [2].Systems for Technology-Enhanced Learning (TEL) need a classification of competencies to automatically react to student actions at the micro-level (e.g. with feedback) and at the macro-level (e.g. with learning objects' (LOs) sequencing), and to build a student model by processing these reactions. Such a classification needs to be reusable, as reproducible as possible, and coherent in order to enable reuse of LOs. In concrete domains it should reflect an optimal granularity to support instructional effectiveness [16] and it should be based on a cognitive task analysis.

show abstract

The Effect of Model Granularity on Student Performance Prediction Using Bayesian Networks

Cited by 27 publications

References 7 publications

Predicting Academic Performance Based on Students’ Blog and Microblog Posts

Predicting Academic Performance Based on Students’ Blog and Microblog Posts

Predicting Students’ Academic Performances – A Learning Analytics Approach using Multiple Linear Regression

Supporting Flexible Competency Frameworks

Contact Info

Product

Resources

About