Data Preprocessing of FRBCS for Early Warning of Student Learning

Qun Zhao1a, Jin-Long Wang2b, Pei-Chen Hung2c*, Shu-Yuan Chuang2d
 
1College of Science & Technology, Ningbo University, Ningbo, China
2Ming Chuan University, Taipei, Taiwan, R.O.C.
Corresponding author: Pei-Chen Hung, Phone number: +886-2-28824564, E-mail: laura@mail.mcu.edu.tw
 
Abstract: 
This paper uses the log data in Moodle system to predict students' learning performance at the early stage of a semester. Since the data quality has great influence on the prediction accuracy, the Normal transformation and the Z transformation are utilized in the preprocessing phase. Then, the Fuzzy Rule- Based Classification System (FRBCS) is employed to create prediction model. The experiment results illustrate that data with Normal distribution can offer the higher prediction accuracy than other methods.
 
Keywords: 

Learning warning, Prediction, Data preprocessing, Fuzzy rule-based classification system

pages: 

425-428

DOI: 

https://doi.org/10.35745/ecei2019v2.108

Year: 

2019

Published in: 

2nd Eurasian Conference on Educational Innovation 2019

Full paper download: 
PDF icon 425-428.pdf