Learning Analytics to identify dropout factors of Computer Science studies through Bayesian networks

Abstract

Student dropout in Engineering Education is an important problem which has been studied from different perspectives, as well as using different techniques. This manuscript describes the methodology used in order to address this question in the context of learning analytics. Bayesian networks have been used as they provide adequate methods for the representation, interpretation and contextualization of data. The proposed approach is illustrated through a case study about Computer Science (CS) dropout at the University of Castilla-La Mancha (Spain), which is close to 40%. To that end, several Bayesian networks were obtained from a database which contained 383 records representing both academic and social data of the students enrolled in the CS degree during four courses. Then, these probabilistic models were interpreted and evaluated. The results obtained revealed that the best model that fits the data is provided by the K2 algorithm although the great heterogeneity of the data studied did not permit the adjustment of the dropout profile of the student too accurately. Nonetheless, the methodology described here can be taken as a reference for future works.

Predicting Computer Engineering students' dropout in Cuban Higher Education with pre-enrollment and early performance data

“…Despite the high classification rate attained, this result is not helpful either for decision makers, as it has a clear correspondence with the promotion regulations of Cuban Higher Education explained in Section 3. Lacave et al (2018) obtain similar results with a smaller sample to predict dropout in Computer Science within the Spanish context. In their study, the previous academic index also has a prominent role.…”

Utilising learning analytics to support study success in higher education: a systematic review

“…While standards for data models and data collection, such as xAPI (Experience API), exist (Kevan and Ryan 2016), learning analytics research and development need to clearly define standards for reliable and valid measures, informative visualisations, and design guidelines for pedagogically effective learning analytics interventions (Seufert et al 2019). In particular, personalised learning environments are increasingly in demand and are valued in higher education institutions for creating tailored learning packages optimised for each individual learner based on their personal profile, containing information such as their geo-and socio-demographic backgrounds (Lacave et al 2018), previous qualifications (Daud et al 2017), their engagement in the recruitment journey (Berg et al 2018), activities on websites (Seidel and Kutieleh 2017), and tracking information on their searches (Macfadyen and Dawson 2012).…”

“…As shown in Table 1, another set of data used for predicting study success is based on students' background information, such as demographics (e.g., age, gender), socio-economic status (e.g., family income, background, expenditure), prior academic experience and performance (Daud et al 2017;Djulovic and Li 2013;Guarrin 2013). For example, Lacave et al (2018) use enrolment age, prior choice of subject and information on scholarships in order to predict student dropouts. In addition to demographic variables (Aulck et al 2017;Sarker 2014), the student's academic self-concept, academic history and workrelated data are used to predict student performance (Mitra and Goldstein 2015), while others use GPA, academic load, and access to counselling (Rogers et al 2014), the student's financial background (Thammasiri et al 2014), or academic performance history (Bydzovska and Popelinsky 2014;Sales et al 2016;Srilekshmi et al 2016) as predictors of students at risk.…”

Automatic Assessment of Students’ Engineering Design Performance Using a Bayesian Network Model

“…Among those common EDM techniques, Bayesian inferential processes such as the application of Bayesian Networks (BNs) are particularly suited to perform certain inferential tasks in assessing student performance in engineering design (Wipulanusat et al., 2020). With relevant factors appropriately included and their dependencies set in terms of conditional probabilities, a BNs model could be seen as a representation of the real-world scenario under consideration (Asif et al., 2017; Lacave et al., 2018; Millán et al., 2010; Ramírez-Noriega et al., 2017). Given the “white-box” nature of BNs that relevant variables and their relationships are carefully worked out to abstract the situation of interest, the posterior probabilities of each affected node could be updated with ready interpretability during the challenge learning process.…”