Welcome to COMSCI 2023

2nd International Conference on Computer Science and Information Technology (COMSCI 2023)

July 15-16, 2023, Virtual Conference

Accepted Papers
Multiclass Classification of Alzheimer’s Disease

Aditya Kumar, Akhila Abhijith, Anmol Kumar, Dr. M Dakshayini, Department of Information Science and Engineering, B.M.S. College of Engineering, Bengaluru, India


Alzheimer’s Disease (AD) is a prevalent .illness in the world generally seen in the age of 65 and older. This disease is generally characterised mainly by loss of memory. Since there is no cure for AD, detecting if the person has AD in early stages can help in the care and diagnosis of the person. Memory problems in AD increase as the disease advances, also developing additional symptoms. We have worked on the AD dataset from Kaggle and AD is classified into 4 categories in our dataset - Normal Brain, Very Mild AD, Minor AD, Modest AD. The art CNN networks not doing a good job in classifying the images and the dataset was divided into training and testing in a way which distorted the model to make wrong predictions. The model we worked on aims to classify MRI images of the brain through a Convolutional Neural Network, useful to detect four stages of dementia in Alzheimer disease’s screening. We started from a robust Convolutional architecture, but the model contained some defects regarding the dataset involved to train it. The problems we detected were the incoherent subdivision of training set and testing set, with totally different sections of the brain in the two sets, the unbalance of the images labelled in the four classes to be classified and the confusion matrix far away from being an identity matrix. We achieved an accuracy of 99.83%.


Alzheimer’s disease, Deep Learning, Deep Convolutional Neural Network, Brain MRI, Multi-Class.

The Development of the Scale of Students' Engagement in Classroom Learning Under the Flipped Classroom Teaching Mode

Yan Jin, Faculty of Psychology and Education, Universiti Malaysia Sabah, Sabah, Malaysia


This study aimed to develop a classroom engagement scale for university students in the 'flipped classroom' model. The participants were 450 students. In developing the items for the scale, a detailed review of the literature and interviews were conducted. Four factors were identified - cognitive engagement, peer relations ( engagement-I ), relationship with the instructor ( emotional engagement-II ), and behavioral engagement - with a total of 21 items. The total Cronbach's alpha coef icient for the scale developed in this way was tested to be 0.959, the cumulative contribution variance in the Exploratory Factor Analysis was 77.970%, and the theoretical model of student engagement in the Confirmatory Factor Analysis fitted the data well. The results show that the scale has good reliability and validity, and can be used to test students' engagement in classroom learning under the 'flipped classroom'.


Student Engagement, Flipped Classroom, Evaluation Scales.

Ddos Detection in Software-defined Network (Sdn) Using Machine Learning

Haya Alubaidan, Reem Alzaher, Maryam AlQhatani, Rami Mohammed, College of Computer Science and Information Technology, Imam Abdulrahman Bin Faisal University, Saudi Arabia


In recent years, the concept of cloud computing and the software-defined network (SDN) have spread widely. The services provided by many sectors such as medicine, education, banking, and transportation are being replaced gradually with cloud-based applications. Consequently, the availability of these services is critical. However, the cloud infrastructure and services are vulnerable to attackers who aim to breach its availability. One of the major threats to any system availability is a Denial-of-Service (DoS) attack, which is intended to deny the legitimate user from accessing cloud resources. The Distributed Denial-of-Service attack (DDoS) is a type of DoS attack which is considerably more effective and dangerous. A lot of efforts have been made by the research community to detect DDoS attacks, however, there is still a need for further efforts in this germane field. In this paper, machine learning techniques are utilized to build a model that can detect DDoS attacks in Software-Defined Networks (SDN). The used ML algorithms have shown high performance in the earliest studies; hence they have been used in this study along with feature selection technique. Therefore, our model utilized these algorithms to detect DDoS attacks in network traffic. The outcome of this experiment shows the impact of feature selection in improving the model performance. Eventually, The Random Forest classifier has achieved the highest accuracy of 0.99 in detecting DDoS attack.


Cloud Computing; Distributed Denial of Service (DDoS); Software-Defined Network (SDN); Machine Learning.