Article
Computer Science, Artificial Intelligence
Joao Gabriel Correa Krueger, Alceu de Souza Britto Jr, Jean Paul Barddal
Summary: School dropout is a global socio-economic problem. Predictive models are developed to determine the likelihood of students dropping out. This paper proposes an approach for creating and enriching a dropout prediction dataset using data from 19 schools in Brazil. Experiments achieved high precision, recall, and KS rates when predicting dropout at different year moments. The study also identifies potential reasons for student dropout.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Vasileios Tsekenis, Charalampos K. Armeniakos, Viktor Nikolaidis, Petros S. Bithas, Athanasios G. Kanatas
Summary: This study aims to utilize artificial intelligence to deploy an integrated sensor system monitoring man overboard incidents during ship voyages, providing accurate and reliable information for prompt detection and rescue of victims.
Article
Computer Science, Information Systems
Pedro Sobreiro, Jose Garcia-Alonso, Domingos Martinho, Javier Berrocal
Summary: Dropout prediction is an important problem that existing approaches fail to address the dynamic nature of dropout risk over time. This study explores the use of random survival forests combined with clusters to improve the prediction performance, and conducts an empirical study among Health Club members. The results show a significant improvement in prediction performance with the model using clusters.
Article
Computer Science, Artificial Intelligence
Zhongshan Chen, Xinning Feng, Shengwei Zhang
Summary: The Internet of Things (IoT) has been widely used in various fields, and its rapid adoption and development have generated a significant amount of evidence. This study proposes an optimized IoT architecture and protocol called FRED-IoT for face recognition and emotion detection in autonomous vehicles. The system aims to track drivers' emotions and facial recognition and improves reliability compared to existing technology, achieving a high rating (F-score) of 96%.
IMAGE AND VISION COMPUTING
(2022)
Article
Mathematics
Diego Opazo, Sebastian Moreno, Eduardo Alvarez-Miranda, Jordi Pereira
Summary: The study found that student dropout is a global issue, and research on two Chilean universities suggests that machine learning models should be applied separately for each university when predicting first-year engineering student dropout rates. The results show that a higher score in almost any entrance university test decreases the probability of dropout, with the mathematical test being the most important variable, while the language test score increases the probability of dropout.
Article
Chemistry, Multidisciplinary
Zihan Song, Sang-Ha Sung, Do-Myung Park, Byung-Kwon Park
Summary: The core of dropout prediction lies in the selection of predictive models and feature tables. Machine learning models have been shown to accurately predict student dropouts. The length of student history data poses a challenge for generating feature tables. Current studies mostly focus on predicting dropouts in the first academic year, assuming that majority of dropouts occur in that year. However, our study based on a dataset from a Korean university reveals that dropouts are evenly distributed across all academic years, emphasizing the importance of dropout prediction for students in any academic year.
APPLIED SCIENCES-BASEL
(2023)
Article
Economics
Marco Delogu, Raffaele Lagravinese, Dimitri Paolini, Giuliano Resce
Summary: Predicting university dropout is crucial for policymaking to prevent dropout and protect national resources and human capital. This study uses machine learning algorithms to predict dropout and finds that random forest and gradient boosting machines show significant predictive capabilities, with first-year academic performance playing a crucial role.
ECONOMIC MODELLING
(2024)
Review
Education & Educational Research
Jing Chen, Bei Fang, Hao Zhang, Xia Xue
Summary: High dropout rates in MOOCs are a common issue, and using machine learning methods for dropout prediction is important. However, current reviews have limitations in terms of lacking a unified definition of dropout, an overall prediction framework, and exploration of key challenges. Therefore, this study proposes three categories of dropout definitions and constructs an overall framework including factors affecting dropout, feature extraction methods, machine learning methods, and evaluation methods. Additionally, key challenges such as interpretability, imbalanced data, and semantic learning trajectory modeling are identified.
INTERACTIVE LEARNING ENVIRONMENTS
(2022)
Article
Computer Science, Information Systems
Sicong Zhou, Huawei Huang, Ruixin Li, Jialiang Liu, Zibin Zheng
Summary: Federated learning (FL) is widely used in IoT applications, but it is vulnerable to various attacks. Traditional defense strategies focus on centralized settings, while decentralized settings require new defensive strategies. Therefore, we propose a committee-based FL system, called ComAvg, which provides a general coordination scheme to address the challenges in decentralized FL.
COMPUTER COMMUNICATIONS
(2023)
Article
Physics, Applied
Vinod Kumar, Neelendra Badal, Rajesh Mishra
Summary: This paper introduces an algorithm for measuring fatigue based on eye-opening, discusses data transmission issues, and highlights the importance of elderly sleep patterns in accurately predicting their health conditions. The use of a machine learning model has proven effective in testing real-world scenarios and accurately predicting conditions of the elderly.
MODERN PHYSICS LETTERS B
(2021)
Article
Chemistry, Analytical
John Oyekan, Windo Hutabarat, Christopher Turner, Ashutosh Tiwari, Hongmei He, Raymon Gompelman
Summary: Cyber-physical systems like satellite telecommunications networks generate large amounts of data, but only a small subset is currently used for troubleshooting, leading to potential long periods of undetected issues. This research proposes a knowledge-based cognitive architecture supported by machine learning algorithms for monitoring satellite network traffic, to assist engineers in identifying and understanding the causes of network faults.
Article
Computer Science, Artificial Intelligence
Geonhee Lee, Jae-Hoon Kim
Summary: The growth of IoT has led to diverse networks, posing a challenge in communication compatibility due to multiple networking interfaces and embedded protocols. To address this, a general-purpose message parser utilizing a recurrent neural network model with stack memory (RNN-SM) is proposed. This parser can extract crucial keywords from various communication network messages and improves keyword extraction accuracy through training on multiple network protocol specifications. The RNN-SM's robust keyword extraction capability makes it an effective solution for facilitating smooth communication and filtering out noise in the IoT.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Education & Educational Research
Kamal Samy Selim, Sahar Saeed Rezk
Summary: Compulsory school dropout is a serious problem that affects both education systems and the overall development of a country. This paper focuses on developing a Logistic classifier to predict students at risk of dropping out early, using an imbalanced Egyptian survey dataset. Different resampling techniques are compared to improve the classifier's performance. The study identifies key factors such as student chronic diseases, co-education, parents' illiteracy, educational performance, and teacher caring, which align with previous research findings in similar countries.
EDUCATION AND INFORMATION TECHNOLOGIES
(2023)
Article
Chemistry, Multidisciplinary
Janka Kabathova, Martin Drlik
Summary: This research focused on predicting student dropout using machine learning classifiers, emphasizing the importance of data understanding and collection, highlighting the limitations of the educational dataset, and demonstrating the performance comparison of several machine learning classifiers.
APPLIED SCIENCES-BASEL
(2021)
Article
Agronomy
Youssef Ahansal, Mourad Bouziani, Reda Yaagoubi, Imane Sebari, Karima Sebari, Lahcen Kenny
Summary: This article provides an overview of the use of UAV, machine learning, and IoT in analyzing crop water status for better irrigation management in arboriculture. The information collected from UAV images and IoT sensors, along with the ability of machine learning models, play a crucial role in improving irrigation efficiency.
Article
Automation & Control Systems
Geetanjali Rathee, Sahil Garg, Georges Kaddoum, Bong Jun Choi, Mohammad Mehedi Hassan, Salman A. AlQahtani
Summary: This article proposes a secure, reliable, and trusted decision-making scheme for collaborative AIoT using multiattribute methods. It uses backpropagation and Bayesian's rule to ensure fast and accurate decisions, and agent-based modeling and population-based modeling trust schemes to compute the legitimacy of the communicating model.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Automation & Control Systems
Ke Wang, Zicong Chen, Mingjia Zhu, Siu-Ming Yiu, Chien-Ming Chen, Mohammad Mehedi Hassan, Stefano Izzo, Giancario Fortino
Summary: Artificial intelligence-driven automation is becoming the technical trend in the new automation era. Convolutional neural network (CNN) technology has been widely used in industrial automation for defect detection and machine vision-driven automation for robot-human tracking. However, the high dependence on neural networks can lead to potential failures in defect detection system.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Engineering, Civil
Kashif Naseer Qureshi, Gwanggil Jeon, Mohammad Mehedi Hassan, Md. Rafiul Hassan, Kuljeet Kaur
Summary: Intelligent Transportation Systems (ITS) have gained popularity due to their smart services, but the increasing number of users has raised concerns over data processing, storage, security, and privacy. To address these concerns, this paper proposes a Blockchain-based Privacy-Preserving Authentication (BPPAU) model that uses smart contracts, access control policies, and on-demand functions to manage data while maintaining user privacy. The model's performance is evaluated through simulation tests analyzing transaction cost, transaction per second, and computational time with various data sizes and block times.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Engineering, Civil
Jia Hu, Kuljeet Kaur, Hui Lin, Xiaoding Wang, Mohammad Mehedi Hassan, Imran Razzak, Mohammad Hammoudeh
Summary: This paper proposes a Transfer Learning based Trajectory Anomaly Detection strategy, named TLTAD, for IoT-empowered Maritime Transportation Systems (IoT-MTS). Experimental results show that TLTAD can accurately detect anomalies in ships' trajectories and reduce model training time.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Information Systems
Dong Zhang, Xiujian Liu, Jun Xia, Zhifan Gao, Heye Zhang, Victor Hugo C. de Albuquerque
Summary: The IoT-based smart healthcare system is of significant importance for the accurate diagnosis of cardiovascular disease. This paper proposes a physics-guided deep learning network that incorporates artificial intelligence techniques and considers the importance of coronary artery features to provide explainable and accurate functional assessment for cardiovascular disease.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Khan Muhammad, Hayat Ullah, Mohammad S. Obaidat, Amin Ullah, Arslan Munir, Muhammad Sajjad, Victor Hugo C. de Albuquerque
Summary: This article proposes an efficient deep-learning-based framework for multiperson salient soccer event recognition in the IoT-enabled FinTech. The framework performs event recognition through frames preprocessing, frame-level discriminative features extraction, and high-level events recognition in soccer videos. The results validate the suitability of the proposed framework for salient event recognition in Nx-IoT environments.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Mathematics
Senthil Kumar Jagatheesaperumal, Snegha Rajkumar, Joshinika Venkatesh Suresh, Abdu H. Gumaei, Noura Alhakbani, Md. Zia Uddin, Mohammad Mehedi Hassan
Summary: In order to promote a healthy lifestyle, individuals need to maintain a balanced diet and engage in customized workouts. A framework is presented in this study to assess an individual's health conditions, allowing people to conveniently evaluate their well-being without consulting a doctor. The framework includes a kit that measures various health indicators and requires minimal effort from nurses.
Article
Mathematics
Mohammad Mehedi Hassan, Mabrook S. AlRakhami, Amerah A. Alabrah, Salman A. AlQahtani
Summary: This study proposes a secure edge-assisted deep learning-based framework for automatic COVID-19 detection, utilizing cloud and edge computing assistance with 5G network and blockchain technologies. The use of edge services in artificial intelligence methods has played a significant role in various applications. DL approaches have been successful in COVID-19 detection using chest X-ray images in cloud and edge computing environments, but they have limitations in training dataset size. To overcome this, the study collects data from different hospitals to train a DL model on a global cloud, integrates the trained models for automatic COVID-19 detection, and retrain them locally at hospitals to improve the model.
Article
Computer Science, Hardware & Architecture
Mohammed Altaf Ahmed, Sara A. Althubiti, Victor Hugo C. de Albuquerque, Marcello Carvalho dos Reis, Chitra Shashidhar, T. Satyanarayana Murthy, E. Laxmi Lydia
Summary: The study introduces a technique called CSOTL-VDCRS for vehicle detection and classification in remote sensing images. It utilizes Mask RCNN for vehicle detection and FWNN for classification. By using CSO as a hyperparameter optimizer, the performance of vehicle detection is enhanced. Experimental results demonstrate the superior performance of the proposed CSOTL-VDCRS technique.
COMPUTERS & ELECTRICAL ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Gui-Bin Bian, Zhang Chen, Zhen Li, Bing-Ting Wei, Wei-Peng Liu, Daniel Santos da Silva, Wan-Qing Wu, Victor Hugo C. de Albuquerque
Summary: Learning surgical skills from trained surgeons can enhance the autonomy of surgical robots and provide appropriate assistance during surgery. This study addresses the challenging issue of the remote center of motion (RCM) constraint, which is often neglected in other works. The proposed method models the implicit constraints of manipulation skills using a probabilistic model to maintain flexibility. It also introduces a novel approach to reconcile the inconsistency between the RCM constraint and surgical task space, improving the generalization of learned skills. Experimental validation demonstrates the effectiveness of the proposed method in a tracking task under the RCM constraint, with a root mean square error exceeding the average for operator demonstrations.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Health Care Sciences & Services
Md. Nazmul Islam, Md. Golam Rabiul Alam, Tasnim Sakib Apon, Md. Zia Uddin, Nasser Allheeib, Alaa Menshawi, Mohammad Mehedi Hassan
Summary: The coronavirus epidemic has spread worldwide causing significant health, financial, and emotional devastation. Therefore, developing a highly accurate AI-based auto-COVID detection system is crucial for healthcare services and the population.
Article
Computer Science, Information Systems
Siyuan Liang, Mengna Xie, Sahil Garg, Georges Kaddoum, Mohammad Mehedi Hassan, Salman A. AlQahtani
Summary: In this paper, a UV positioning system REW_SLAM based on lidar and stereo camera is proposed, which achieves real-time online pose estimation of UV by using high-precision lidar pose correction visual positioning data. A six-element extended Kalman filter (6-element EKF) is proposed to fusion lidar and stereo camera sensors information, effectively improving the accuracy of data fusion. Meanwhile, a modified wavelet denoising method is introduced to preprocess the original lidar data to improve its quality. Experimental results show that compared with the other two algorithms, the relative pose error and absolute trajectory error of this algorithm increased by 0.26 m and 2.36 m on average, respectively, while the CPU occupancy rate increased by 6.685% on average, thus proving the robustness and effectiveness of the algorithm.
HUMAN-CENTRIC COMPUTING AND INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Gui-Bin Bian, Jia-Ying Zheng, Zhen Li, Jie Wang, Pan Fu, Chen Xin, Daniel Santos da Silva, Wan-Qing Wu, Victor Hugo C. De Albuquerque
Summary: This study proposes a multimodal, multi-timescale data fusion network based on deep learning to improve the accuracy of continuous circular capsulorhexis (CCC) procedures. Through validation on an ophthalmologist CCC multimodal maneuver dataset, the model demonstrates superior performance in continuous action sequence segmentation and minority class recognition.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Computer Science, Information Systems
Rabeya Khatun Muna, Muhammad Iqbal Hossain, Md. Golam Rabiul Alam, Mohammad Mehedi Hassan, Michele Ianni, Giancarlo Fortino
Summary: This research aims to detect large-scale attacks on IoT devices using the Extreme Gradient Boosting (XG-Boost) classifier and Explainable Artificial Intelligence (XAI) approaches. The results demonstrate that the proposed model can efficiently identify malicious attacks and threats, reducing IoT cybersecurity threats in smart cities.
INTERNET OF THINGS
(2023)