Hybrid Fuzzy-Statistical System for Learning Analytics
- Resource Type
- Conference
- Authors
- Sohail, Shaleeza; Khanum, Aasia; Alvi, Atif
- Source
- 2018 IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE) Teaching, Assessment, and Learning for Engineering (TALE), 2018 IEEE International Conference on. :989-994 Dec, 2018
- Subject
- Engineering Profession
General Topics for Engineers
Predictive models
Education
Analytical models
Data models
Correlation
Machine learning
Expert systems
Learning analytics
Fuzzy Logic
hybrid system
- Language
- ISSN
- 2470-6698
Learning analytics, being a very active field of research, recently has a number of significant contributions where data related to students' demographics, learning and behaviour is being used to predict student performance. In higher education institutes most of these models are based on statistical and machine learning techniques. The biggest challenge that can be seen in almost all these research efforts is the ineffective portability of prediction model across multiple cohorts even in the same institute. We propose to use hybrid fuzzy-statistical model for predicting student performance based on behaviour related predictors in a particular cohort. We have recently started this project so preliminary details of our model based on Fuzzy Expert System are provided. The model is proposed for enhancing traditional statistical models with Fuzzy Sets/Fuzzy Logic to overcome the limitations of commonly used statistical techniques and deliver a stronger system with combined advantages of both domains.