SVM and PCA Based Learning Feature Classification Approaches for E-Learning System

SVM and PCA Based Learning Feature Classification Approaches for E-Learning System

Aditya Khamparia, Babita Pandey
DOI: 10.4018/IJWLTT.2018040103
(Individual Articles)
No Current Special Offers


E-learning and online education has made great improvements in the recent past. It has shifted the teaching paradigm from conventional classroom learning to dynamic web based learning. Due to this, a dynamic learning material has been delivered to learners, instead of static content, according to their skills, needs and preferences. In this article, the authors have classified eight different types of student learning attributes based on National Centre for Biotechnical Information (NCBI) e-learning database. The eight types of attributes are Anxiety (A), Personality (P), Learning style (L), Cognitive style (C), Grades from previous sem (GP), Motivation (M), Study level (SL) and Student prior knowledge (SPK). In this article the authors have proposed an approach which uses principal components of student learning attributes and have later independently classified these attributes using feed forward neural network (NN) and Least Square –Support Vector Machine (LS-SVM).
Article Preview

1. Introduction

In order to compete and survive in the twenty-first century internet based economy, it is essential that students are aware of learning programming techniques and acquire these skills in their behaviour (Chen, Warden, & Chang, 2006). The goal of instructing programming is to facilitate students’ future by helping them build skills to improve logic to solve real world problem so that they become successful in their academic and professional career. In the conventional classroom learning system instructors deliver the same content to all students without taking into consideration their individuals personality type, learning style, anxiety level, prior knowledge on programming, cognitive styles or result of previous grades. Such type of current learning is based on static learning material but not on dynamic programming based learning (Romero and Bra, 2007). In such type of system if students want to maximize their learning outcome, it is they who have to adapt according to the course content and cannot change the course content to deliver based on their specific attributes or characteristics.

In this paper, our intent is to classify student learning attributes with the help of principal component analysis. Intelligent computing techniques like neural network, principal component analysis and support vector machines are used to investigate students’ background and their attributes or characteristics in order to optimize their learning sequence and maximize their learning outcome in computer programming courses.

An online web based system was designed to analyze various types of student attributes (Gender (G), Personality (P), Anxiety (A), Learning level (L) and Cognitive ability (C)) using Case based reasoning, hybrid neural network and efficient data mining technique in learning C programming course. This system yielded an average accuracy and sensitivity of 71.65% and 70.22% respectively with five different classes (Khamparia and Pandey, 2015). An English language learning system was designed along with mainly three student features (Gender, Personality and Anxiety) as a classifier which mainly used neural network and decision tree algorithm. Gender, study level, Student prior knowledge and intrinsic motivation were classified using hybrid neural network with post-test classification and yielded an accuracy of 67% (Van Seters et al. 2012). Two types of student attributes, cognitive style and learning style, were classified using particle swarm optimization with neural network classifier and achieved more than 64% of sensitivity and 65% of accuracy (Yang et al. 2008).

Even though many works were reported on students’ attribute classification, there is a constant need to improve the classification accuracy when more number of attributes are incorporated and operated upon large databases comprising of different instances of individual attributes. Most of the proposed methods involve too many computational overhead, but our intent is to develop a simple methodology for students’ learning attribute classification with efficient accuracy for large instances of e-learning database.

The paper is organized in the following ways: Section 2 represents the literature review. It also represents the proposed methodology with involvement of Principal Component Analysis (PCA) and classifiers. Section 3 represents the results and computing description. We present the interpretation of result in Section 4. Finally, we conclude the paper in Section 5.

Complete Article List

Search this Journal:
Volume 19: 1 Issue (2024)
Volume 18: 2 Issues (2023)
Volume 17: 8 Issues (2022)
Volume 16: 6 Issues (2021)
Volume 15: 4 Issues (2020)
Volume 14: 4 Issues (2019)
Volume 13: 4 Issues (2018)
Volume 12: 4 Issues (2017)
Volume 11: 4 Issues (2016)
Volume 10: 4 Issues (2015)
Volume 9: 4 Issues (2014)
Volume 8: 4 Issues (2013)
Volume 7: 4 Issues (2012)
Volume 6: 4 Issues (2011)
Volume 5: 4 Issues (2010)
Volume 4: 4 Issues (2009)
Volume 3: 4 Issues (2008)
Volume 2: 4 Issues (2007)
Volume 1: 4 Issues (2006)
View Complete Journal Contents Listing