Employing Data Analytics for Academic Improvement of Students
- Resource Type
- Conference
- Authors
- Pathak, Rahul; Beig, Mirza Tanweer; Faraz, Saquib; Jain, Mayank
- Source
- 2019 International Conference on Machine Learning, Big Data, Cloud and Parallel Computing (COMITCon) Machine Learning, Big Data, Cloud and Parallel Computing (COMITCon), 2019 International Conference on. :385-389 Feb, 2019
- Subject
- Communication, Networking and Broadcast Technologies
Computing and Processing
Predictive models
Mathematical model
Data models
Measurement
Education
Biological system modeling
Linear regression
Logistic regression
Linear Regression
VIF values
Kappa metric
Confusion matrix.
- Language
Mathematical models which can aid academic assessment of students have the potential to improve the delivery of education and learning. Such models can be utilized to find out the quality of learning and influencing factors. These insights enable stakeholders to take effective measures for the attainment of educational objectives. Present work attempts to formulate two mathematical models for assessment: a predictor model for final grade predictions of students and a classifier model to identify academically weak students. Linear regression was employed for predictions while logistic regression was used for the classification. The Computation was performed in R environment. The study dealt with a dataset containing a record of three hundred ninety-five students with thirty-three attributes. Predictor model was tested with statistical metrics like VIF value, fit chart, and residual plots. Performance of classifier was investigated by using confusion matrix and kappa value. Predictor model gave a very low average error in score prediction of final grades. Classifier model demonstrated high accuracy of eighty-eight percent in identifying weak students on the test sample.