In the online education field, Massive open online courses (MOOCs) have become popular in recent years. Educational institutions and Universities provide a variety of specialized online courses that helps the students to adapt with various needs and learning preferences. Because of this, institutional repositories creates and preserve a lot of data about students' demographics, behavioral trends, and academic achievement every day. Moreover, a significant problem impeding their future advancement is the high dropout rate. For solving this problem, the dropout rate is predicted by proposing an Ensemble Deep Learning Network (EDLN) model depending on the behavior data characteristics of learners. The local features are extracted by using ResNet-50 and then a kernel strategy is used for building feature relations. After feature extraction, the high-dimensional vector features are sent to a Faster RCNN for obtaining the vector representation that incorporates time series data. Then an attention weight is obtained for each dimension by applying a static attention mechanism to the vector. Extensive experiments on a public data set have shown that the proposed model can achieve comparable results with other dropout prediction methods in terms of precision, recall, F1 score, and accuracy.
1 March 2023 in Indonesian Journal of Electrical Engineering and Computer Science
Information and communication technologies (ICTs) in all aspects of life have become the tools that are present everywhere and in a ubiquitous manner. Over the last twenty to thirty years, it has been noticed that the application of ICT has significantly changed the procedures and courses of almost all the higher education institutions. The higher education field is a highly socially-focused practice and providing quality education is traditionally been linked with the efficient teacher and their capacity for a high level of one-on-one contact with students. The use of ICT in higher education allows for more student-centered learning settings. Due to the reason that the world is quickly developing into a digital place with a huge level of digital information, however, the role of ICT in higher education is becoming increasingly significant and will persistently to grow and evolve in the 21st century. The application of ICT in the teaching and learning process, as well as its effective usage in higher education, depends on a variety of tools and techniques. In this paper, a detailed analysis of various modes adopted by the institutions of higher education is made.
The chapter explores the implications of deep learning in medical sciences, focusing on deep learning concerning natural language processing, computer vision, reinforcement learning, big data, and blockchain influence on some areas of medicine and construction of end-to-end systems with the help of these computational techniques. The deliberation of computer vision in the study is mainly concerned with medical imaging and further usage of natural language processing to spheres such as electronic wellbeing record data. Application of deep learning in genetic mapping and DNA sequencing termed as genomics and implications of reinforcement learning about surgeries assisted by robots are also overviewed. Request access from your librarian to read this chapter's full text.
Lung cancer remains a leading cause of cancer-related deaths, primarily due to late-stage detection. Although medical imaging and biopsy-based evaluations have improved, early identification of lung cancer continues to be challenging. To address this, we propose a Vision Transformer (ViT)-based model for binary lung nodule classification using computed tomography (CT) images. This study uses a Kaggle-hosted subset of the LIDC–IDRI dataset containing 315 CT nodule patches, where the original malignancy scores were converted into benign and malignant binary classes. Given the small dataset size, an extensive augmentation pipeline was designed to enhance model generalization. The lightweight ViT-Small/16 architecture demonstrated strong performance, achieving 92.3% accuracy, 90.5% precision, 93.8% recall, and a 92.1% F1-score. These results highlight the potential of compact transformer models for early lung cancer identification. This work is among the first to evaluate ViT-Small/16 on a small-scale CT nodule dataset using a tailored augmentation strategy for limited-data medical imaging.
1 March 2023 in Indonesian Journal of Electrical Engineering and Computer Science
The efficient finding of common patterns: a group of items that appear frequently in a dataset is a critical task in data mining, especially in transaction datasets. The goal of this paper is to look into the efficiency of various algorithms for frequent pattern mining in terms of computing time and memory consumption, as well as the problem of how to apply the algorithms to different datasets. In this paper, the algorithms investigated for mining the frequent patterns are; Pre-post, Pre-post+, FIN, H-mine, R-Elim, and estDec+ algorithms. These algorithms have been implemented and tested on four real-life datasets that are: The retail dataset, the Accidents dataset, the Chess dataset, and the Mushrooms dataset. From the results, it has been observed that, for the Retail dataset, estDec+ algorithm is the fastest among all algorithms in terms of run time as well as consumes less memory for its execution. Pre-post+ algorithm performs better than all other algorithms in terms of run time and maximum memory for the Mushrooms dataset. Pre-Post outperforms other algorithms in terms of performance. And for Accident datasets, in terms of execution time and memory consumption, the FIN method outperforms other algorithms.