• PDF / 1,348,210 Bytes
• 16 Pages / 439.37 x 666.142 pts Page_size
• 4 Downloads / 198 Views

DOWNLOAD

REPORT

---

Hybridization of cluster-based LDA and ANN for student performance prediction and comments evaluation Sakshi Sood 1

& Munish

Saini 1

Received: 23 June 2020 / Accepted: 1 November 2020/ # Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract For a productive life, education plays a critical role to fill individual life with value and excellence. Education is compulsory to provide things that individuals partake in to compete in the modern world. Predicting the academic performance of the student is the most successive research in this era. A different set of approaches and methods are incorporated to increase student performance. However, this is a challenging task due to the wrong course selection. In the proposed study, we have used the hybrid approach consisting of Cluster-based Linear Discriminant Analysis (CLDA) and Artificial Neural Network (ANN) to provide the prospective students with the motivational comments and the video recommendations by which students can choose the right subject and the comments will facilitate the students with the insight reasons of dropout opted by other students for this course. The outcomes of this study will help in the reduction of the number of dropouts. The students will be able to choose an appropriate course for performance enhancement and carrier excel. Keywords Cluster-based linear discriminant analysis (CLDA) . Student performance .

Dropouts . Classification . Prediction . Machine learning

1 Introduction Data mining is one of the most cardinal areas in recent technologies for retrieving valid information from the huge amount of unstructured and distributed data using parallel processing of data (Haraty et al. 2015). Data mining techniques are applied in various fields to find novel information from the huge data set (Mohan et al. 2018). Recently,

* Sakshi Sood [email protected]

1

Department of Computer Engineering and Technology, Guru Nanak Dev University, Amritsar, Punjab, India

Education and Information Technologies

data mining technology has been used in the educational filed to extract the hidden information from educational data sets. It also supports the classical educational system facilitating teachers to analyze what students know and what learning techniques are most effective for students. Educational data mining and learning analytics employ technologies from statistics, computer science, and machine learning to extract useful information from collected educational data, gain valuable insight into learning and find out solutions to improve learning performance and teaching effectiveness. Nowadays, the digitization is used by the universities in teaching-learning and other academic processes causing the generation of a huge amount of digital data. This data is helpful for teachers, policymakers, and administrators for decision making if it is effectively transformed. By providing timely information to different stakeholders, it advances the quality of educational processes. It is found that in many countries (including dev