Multidimensional Approach Based on Deep Learning to Improve the Prediction Performance of DNN Models

Authors

  • Mohamed El Fouki Abdelmalek Essaadi University, Tetouan, Morocco
  • Noura Aknin Abdelmalek Essaadi University, Tetouan, Morocco
  • Kamal Eddine El Kadiri Abdelmalek Essaadi University, Tetouan, Morocco

DOI:

https://doi.org/10.3991/ijet.v14i02.8873

Keywords:

Educational Data Mining (EDM), Classification, Deep Neural Network (DNN), Deep Learning, Principal Component Analysis (PCA)

Abstract


The most of collected data samples from E-learning systems consist of correlated information caused by overlapping input instances, which decrease the classifier credibility and reliability. This paper presents an improved classification model based on Deep Learning and Principal Component Analysis (PCA) method as its use in reducing the dimensionality of data. By this task, we introduce the best learning process to extract just the useful parameters that describe students’ per-formances in an E-learning system. One of the primary goals of this technique is to help earlier in detecting the dropouts and discovering of students who need special attention, so that the teachers could provide the appropriate counseling at the right time. This study presents the proposal approach and its algorithms. In addition, it shows how deep neural network was modeled in the training phase, and how PCA helps in the elimination of correlated information in our dataset to increase the classifier performance. Finally, we introduce an example of an appli-cation of the method in a data mining scenario, find out more references for fur-ther information.

Author Biographies

Mohamed El Fouki, Abdelmalek Essaadi University, Tetouan, Morocco

Mohammed EL FOUKI received the Master degree in Computer Science in 2014 from Abdelmalek Essaadi University in Tetuan, Morocco. Currently, he is a PhD Candidate and member of Computer Science, Operational Research and Applied Statistics Laboratory in the same university.

Noura Aknin, Abdelmalek Essaadi University, Tetouan, Morocco

Noura Aknin , Professor of Electrical & Computer Engineering at Abdelmalek Es-saadi University since 2000. She received PhD degree in Electrical Engineering in 1998. She is the Head of Research Unit Information Technology and Modeling Sys-tems. She is the Co-founder of the IEEE Morocco Section since November 2004 and a member of several IEEE societies. Noura AKNIN is R&D project manager/member related to new technologies and their applications. She was a chair of several confer-ences and she has been involved in the organizing and in the Scientific Committees of several international conferences held worldwide dealing with computer science and applications

Kamal Eddine El Kadiri, Abdelmalek Essaadi University, Tetouan, Morocco

Kamal Eddine El Kadiri is a Professor of computer science and mathematics at Abdelmalek Essaadi University, Morocco, since 1985. He received the “Graduate thesis” degree in Data analysis at Paris VI University in 1984 and “Thesis status” degree in Datamining and Computer Science from Granada in 1994. Currently, he is the Director of Computer Science, Operational Research and Applied Statistics La-boratory since 2007. Kamal Eddine EL KADIRI is responsible of several R&D pro-jects and member of several International Research projects. He is also part of many boards of international journals and international conferences.

Downloads

Published

2019-01-30

How to Cite

El Fouki, M., Aknin, N., & El Kadiri, K. E. (2019). Multidimensional Approach Based on Deep Learning to Improve the Prediction Performance of DNN Models. International Journal of Emerging Technologies in Learning (iJET), 14(02), pp. 30–41. https://doi.org/10.3991/ijet.v14i02.8873

Issue

Section

Papers