Optimization of Convolutional Neural Network Architecture with Genetic Algorithm for Feature Extraction Prediction of Student Graduation


Authors : Abi Mestu Yudansha; M Arief Soeleman; Affandy

Volume/Issue : Volume 9 - 2024, Issue 2 - February

Google Scholar : https://tinyurl.com/ydrfnjcz

Scribd : https://tinyurl.com/5cm7y62c

DOI : https://doi.org/10.5281/zenodo.10784755

Abstract : Higher education institutions such as universities play a central role in the steps where comprehensive research and development activities take place within a highly competitive environment. The academic achievements of students become a crucial element within the structure of these higher education institutions. This is because one of the key indicators of university quality is an exceptional track record of academic achievements. Universitas Dian Nuswantoro (UDINUS), a private educational institution with an A accreditation rating, is located in Semarang, Indonesia. One of the faculties that holds a significant role at UDINUS is the Faculty of Computer Science, which stands out with the highest number of students, particularly in the Bachelor's program of Computer Science (S1), which records a comprehensive and outstanding student count compared to other study programs. Therefore, it is appropriate to focus research on the data regarding the graduation rates of students from the Computer Science S1 program. In this study, the author applies Data Mining, a method involving manipulation of large-scale data. The primary mission of this research is to address the question of how the implementation of Deep Learning using an optimized Convolutional Neural Network (CNN) through Genetic Algorithm can be utilized to predict student graduation. Consequently, the outcomes can serve as references to expedite student graduation. The study demonstrates that the feature extraction values using CNN and the hyperparameters using Genetic Algorithm show an overall increase in accuracy when using K-Nearest Neighbor (K-NN) for all values of n: 3, 4, 5, and 6 (Proven that feature extraction from tabular data represented as images and processed with the CNN algorithm using the most suitable parameters is successful)

Keywords : Component Udinus, Convolutional Neural Network (CNN), Genetic Algorithm, K- Nearest Neighbor (KNN).

Higher education institutions such as universities play a central role in the steps where comprehensive research and development activities take place within a highly competitive environment. The academic achievements of students become a crucial element within the structure of these higher education institutions. This is because one of the key indicators of university quality is an exceptional track record of academic achievements. Universitas Dian Nuswantoro (UDINUS), a private educational institution with an A accreditation rating, is located in Semarang, Indonesia. One of the faculties that holds a significant role at UDINUS is the Faculty of Computer Science, which stands out with the highest number of students, particularly in the Bachelor's program of Computer Science (S1), which records a comprehensive and outstanding student count compared to other study programs. Therefore, it is appropriate to focus research on the data regarding the graduation rates of students from the Computer Science S1 program. In this study, the author applies Data Mining, a method involving manipulation of large-scale data. The primary mission of this research is to address the question of how the implementation of Deep Learning using an optimized Convolutional Neural Network (CNN) through Genetic Algorithm can be utilized to predict student graduation. Consequently, the outcomes can serve as references to expedite student graduation. The study demonstrates that the feature extraction values using CNN and the hyperparameters using Genetic Algorithm show an overall increase in accuracy when using K-Nearest Neighbor (K-NN) for all values of n: 3, 4, 5, and 6 (Proven that feature extraction from tabular data represented as images and processed with the CNN algorithm using the most suitable parameters is successful)

Keywords : Component Udinus, Convolutional Neural Network (CNN), Genetic Algorithm, K- Nearest Neighbor (KNN).

Never miss an update from Papermashup

Get notified about the latest tutorials and downloads.

Subscribe by Email

Get alerts directly into your inbox after each post and stay updated.
Subscribe
OR

Subscribe by RSS

Add our RSS to your feedreader to get regular updates from us.
Subscribe