10th International Congress on Information and Communication Technology in concurrent with ICT Excellence Awards (ICICT 2025) will be held at London, United Kingdom | February 18 - 21 2025.
Sign up or log in to bookmark your favorites and sync them to your phone or calendar.
Authors - Tajim Md. Niamat Ullah Akhund, Kenbu Teramoto Abstract - The demand for efficient human activity recognition systems has surged recently, driven by the need for intelligent monitoring in various environments such as smart homes and workplaces. This paper presents a novel approach to measuring human activeness using a single Passive Infrared (PIR) sensor, highlighting its simplicity, costeffectiveness, and privacy-conscious design. This paper introduces a novel one-dimensional modeling approach for measuring human activeness using a single Passive Infrared (PIR) sensor, incorporating the Laplace distribution to analyze movement patterns. We define an activeness index μ, quantifying average human activity over time, allowing for precise numerical assessment. Our method utilizes the sensor’s capabilities to gather data on human movement and generate numerical metrics of average activeness over time. The results demonstrate that this approach effectively captures human activity levels while minimizing equipment complexity. This work contributes to the growing field of human activity recognition by offering a practical solution that balances performance with user privacy and affordability.
Authors - Hasti Vakani, Mithil Mistry, Hardikkumar Jayswal, Nilesh Dubey, Nitika Sharma,Rohan Patel, Dipika Damodar Abstract - Obesity has become a significant global health concern due to its as-sociation with various non-communicable diseases. Traditional methods for obesity assessment, such as BMI, often fail to capture the complexity of the condition, highlighting the need for more accurate predictive tools. This research utilize the machine learning algorithms, including Random Forest, Gradient Boosting, Support Vector Machines, and Neural Networks, in a stacking ensemble model to predict obesity levels. Utilizing datasets from diverse populations, the model achieved a high accuracy of 96.69%. Key features such as BMI, age, and dietary habits were identified as critical predictors through Recursive Feature Elimination. The research findings demonstrate the potential of advanced data-driven techniques in providing personalized insights into obesity management and underscore the transformative role of machine learning in public health initiatives.
Authors - Kutub Thakur, Md Liakat Ali, Suzanna Schmeelk, Joan Debello, Denise Dragos Abstract - The escalating prevalence of obesity in young adults has become a pressing public health concern, requiring innovative risk prediction and intervention approaches. This paper examines the potential of combining traditional lifestyle factors with social media behavior to predict obesity risk in young adults while addressing ethical considerations related to data privacy and informed consent. By identifying the most predictive social media metrics associated with obesity risk, this research offers novel insights that could inform targeted prevention strategies. Through a mixed-methods approach, the study examines the associations between social media behavior, traditional lifestyle factors, and obesity risk while ensuring adherence to ethical guidelines and protecting individual privacy. The findings highlight the importance of integrating social media metrics into risk prediction models, offering new avenues for intervention and prevention efforts. This research provides a deeper understanding of the complex interplay between social media behavior, lifestyle factors, and obesity risk, emphasizing the need for multidisciplinary approaches to tackle this growing public health challenge.
Authors - Alisher Ikramov, Shakhnoza Mukhtarova, Raisa Trigulova, Dilnoza Alimova, Dilafruz Akhmedova Abstract - Hospital readmissions pose a significant burden on healthcare systems, especially for patients with type 2 diabetes mellitus (T2DM) and cardiovascular diseases. Early readmission risk prediction is crucial for improving patient outcomes and reducing costs. In this study, we develop a predictive model based on accessible clinical features to estimate the risk of future hospitalizations. Using data from 260 patients at the Republican Specialized Scientific and Practical Medical Center for Cardiology in Uzbekistan, we trained a Generalized Linear Model that achieved a ROC AUC of 0.898 on the test set.
Authors - Lucas V. Santos, Vitor B. Souza Abstract - Fog computing emerges as an innovative solution for edge data processing, proving to be particularly important in the context of the Internet of Things (IoT) by delivering low latency and high bandwidth at the cost of requiring a stable connection. One application that has greatly benefited from this concept is the use of Unmanned Aerial Vehicles (UAVs), also known as drones, for various applications requiring real-time communication between these devices and, potentially, a base station. This paper focuses on the use of UAVs, highlighting the connectivity challenges posed by the limitations of wireless communication technologies, such as Wi-Fi. To address these challenges, we propose a model based on deep reinforcement learning (DDQN), which helps drones make decisions on the best route between the origin and destination, balancing the minimization of travel time and the maximization of connectivity throughout the journey. Using a simulated environment where drones are trained to avoid disconnection areas, we found that the proposed model significantly improves connection stability in areas with limited coverage, albeit with an unavoidable increase in route distance. Comparisons with traditional routing methods demonstrate the advantages of our model.
Authors - Sacrificio Sithole Junior, Mohammad Gulam Lorgat Abstract - The increase in the number of university students has resulted in long queues and delays in services, both during orientation events and in resolving general queries. A service chatbot is an artificial intelligence tool designed to interact with users, answering frequently asked questions and assisting in solving problems in an automated and efficient manner. This study presents the development of a chatbot prototype for the Faculty of Engineering administrative office in Chimoio, at the Universidade Católica de Moçambique (UCM), aiming to optimise service delivery, reduce waiting times, and increase efficiency in resolving common issues. Using a mixed-method approach, the study involved direct observation and questionnaires administered to students to identify the main problems with traditional service. The chatbot's development was carried out in two phases: the first involved data collection and the identification of needs, while the second covered the implementation of the prototype. This chatbot can provide a viable and effective solution to the challenges faced, delivering faster and more efficient service, while freeing up human resources for more complex tasks.
