Learning Management System with Prediction Model and Course-content Recommendation Module

Digna Sayco Evale
Journal of Information Technology Education: Research  •  Volume 16  •  2017  •  pp. 437-457
Aim/Purpose: This study is an attempt to enhance the existing learning management systems today through the integration of technology, particularly with educational data mining and recommendation systems.

Background: It utilized five-year historical data to find patterns for predicting student performance in Java Programming to generate appropriate course-content recommendations for the students based on their predicted performance.

Methodology : The author used two models for the system development: these are the Fayyad knowledge discovery in databases (KDD) process model for the data mining phase and the evolutionary prototyping for system development. WEKKA and SPSS were used to find meaningful patterns in the historical data, while Ruby on Rails platform was used to develop the software.

Contribution: The contribution of this study is the development of an LMS architecture that can be used to augment the capabilities of the existing systems by integrating a data mining technique for modelling the leaners profile; developing of an algorithm for generating predictions; and making the most appropriate recommendations for the learners based on prior knowledge and learning styles.

Findings: The result shows that J48 was the best data mining algorithm to be implemented for finding patterns in the data sets used in this study. Attributes such as age, gender, class schedule, and grades in other programming subjects were found relevant in predicting student performance in Java.

Recommendations for Practitioners : It is recommended that collaboration between the academe and IT industry be strengthened to develop a more advanced LMS which could enhance classroom teaching and improve the learning process.

Recommendation for Researchers: Combination of multiple algorithm in classifying data set is recommended to further improve the algorithm and rule sets of prediction. Inclusion of intrinsic attributes as part of data set aside from personal and academic records is also recommended.

Impact on Society : This LMS can be used to produce independent learners.

Future Research: Study about the impact of implementing this LMS in classroom environment will be conducted on the second phase.
learning management systems, educational data mining, prediction model, per-formance prediction, attribute selection, course-content recommendation, index of learning styles
7 total downloads
Share this
 Back

Back to Top ↑