Publication Search

Excessive sugar consumption has become a serious public health problem. Increasing patterns of food and drink consumption in line with changes in modern lifestyles have contributed to an increase in the prevalence of non-communicable diseases such as obesity, type 2 diabetes and cardiovascular disorders. This study analyzes and analyzes the use of Artificial Intelligence (AI), especially Deep Learning techniques and Neural Network algorithms, in the classification of sugar content in sweetened drinks. The Systematic Literature Review (SLR) method was used to filter relevant studies published between 2020-2024. The study results show that AI is able to provide more efficient and accurate solutions than manual methods. However, although the literature results show great potential, the application of AI in sugar content classification still requires further empirical research. This study emphasizes the importance of developing AI models tailored to the characteristics of sweetened drinks to support consumer decision making regarding healthier drink choices.

This work explores the potential of PennyLane and variational quantum-classical algorithms (VQCA) to forecast lung cancer using a structured dataset. The VQCA model performs exceptionally well, with flawless training, validation, and test accuracies of 1.0, demonstrating its capacity to identify patterns in the dataset and provide reliable predictions successfully. Contrarily, the accuracy of the quantum neural network (QNN) and classical neural network (NN) models is lower, demonstrating the benefits of utilizing quantum computing methods for enhanced predictive modeling. We provide a complete examination of the data, stressing the better performance of the VQCA model and its promise in correctly predicting lung cancer. The results highlight the importance of quantum-classical algorithms and help us understand the benefits and drawbacks of various strategies for predicting lung cancer. The study highlights the potential applications of quantum computing techniques in advancing the field of healthcare analytics. It shows the capability of the VQCA model to predict lung cancer using a tabular dataset accurately. Further research in this area is needed to explore scalability and practical implementation aspects. In summary, this study showcases the potential of VQCA and PennyLane in predicting lung cancer and underscores the benefits of quantum computing techniques in healthcare analytics.

Credit approval prediction is one of the critical challenges in the financial industry, where the accuracy and efficiency of credit decision-making can significantly affect business risk. This study proposes an outlier detection method using the Gaussian Mixture Model (GMM) combined with Extreme Gradient Boosting (XGBoost) to improve prediction accuracy. GMM is used to detect outliers with a probabilistic approach, allowing for finer-grained anomaly identification compared to distance- or density-based methods. Furthermore, the data cleaned through GMM is processed using XGBoost, a decision tree-based boosting algorithm that efficiently handles complex datasets. This study compares the performance of XGBoost with various outlier detection methods, such as LOF, CBLOF, DBSCAN, IF, and K-Means, as well as various other classification algorithms based on machine learning and deep learning. Experimental results show that the combination of GMM and XGBoost provides the best performance with an accuracy of 95.493%, a recall of 91.650%, and an AUC of 95.145%, outperforming other models in the context of credit approval prediction on an imbalanced dataset. The proposed method has been proven to reduce prediction errors and improve the model's reliability in detecting eligible credit applications.

The global clean water crisis is exacerbated by significant losses in water distribution networks (WDNs), resulting in inefficient use of both water and energy resources. Traditional methods of leak detection and pressure management often fail to address these inefficiencies, leading to substantial water wastage and high operational costs. This research aims to design a sustainable, smart water distribution system using advanced technologies such as Machine Learning (ML) for leak detection and automated pressure control. The system employs real-time monitoring through IoT sensors, which continuously gather data on water pressure, flow rates, and other critical parameters. This data is analyzed using various ML algorithms, including supervised and unsupervised learning models, to detect anomalies indicative of leaks. Additionally, the system integrates automated pressure control mechanisms that dynamically adjust pressure to prevent over-pressurization, reducing both water loss and energy consumption. By combining leak detection and pressure control, the proposed system offers a more efficient, sustainable solution to water resource management compared to traditional methods. The expected outcomes include a significant reduction in water loss, enhanced energy efficiency, and improved water service quality. However, the implementation of such a system in rural or small-town infrastructure faces challenges, including sensor maintenance, algorithm reliability, and regulatory issues. A cost-benefit analysis suggests that while the initial investment in smart technologies may be high, the long-term savings in water and energy costs outweigh these costs. This study underscores the potential of ML-based systems in enhancing water conservation, operational efficiency, and sustainability in water management.

The rapid advancement of modern computing has driven extensive research on numerical algorithms for solving large-scale systems of linear equations. Classical methods such as LU decomposition, Jacobi, and Gauss–Seidel have been revisited and optimized to leverage parallel architectures, GPUs, and even quantum platforms. Recent studies demonstrate that optimized algorithms can reduce computation time by more than 50% while maintaining high accuracy in solving high-dimensional problems. LU decomposition, particularly in its parallel and GPU-based implementations, has shown superior performance in batch processing and industrial-scale simulations. Meanwhile, iterative methods such as Jacobi and Gauss–Seidel remain relevant due to their flexibility in numerical modeling, with further developments for block matrix systems, finite element applications, and FPGA architectures. The integration of these enhanced algorithms is not only beneficial for the advancement of scientific software development but also supports practical applications in engineering simulations, large-scale data optimization, and machine learning. Therefore, an integrative review of modern numerical algorithm developments is crucial in bridging the gap between industrial demands and research progress in scientific computing.

Customer churn prediction models have become an important tool in the telecommunications industry to reduce churn rates and improve customer retention. This research focuses on building an accurate customer churn prediction model using machine learning algorithms for TELCO Company. By applying diverse feature engineering techniques and prediction models such as RandomForestClassifier, DecisionTreeClassifier, and XGBoost, this study showcases a significant improvement in prediction accuracy compared to previously implemented rule-based methods. The findings of this research allow TELCO Company to identify high-risk customers more effectively and implement targeted retention strategies. Results show that the resulting model can identify customers at risk of churn more effectively, enabling more targeted retention actions..

In the era of data revolution and artificial intelligence, machine learning model optimization has become one of the most dynamic and crucial research areas. This article reviews the latest techniques in machine learning model optimization with a focus on pursuing maximum performance. We discuss various methods applied to improve model accuracy and efficiency, ranging from hyperparameter tuning techniques, advanced optimization algorithms such as Bayesian optimization, to innovative approaches such as meta-learning and transfer learning. These optimization techniques not only aim to improve model performance but also to overcome challenges related to big data, model complexity, and computational limitations. We investigate how these methods can be integrated in machine learning pipelines to achieve better results with more efficient resources. Through a review of recent literature and case studies of applications in various domains, this article provides in-depth insights into the trends and developments in model optimization, as well as practical recommendations for researchers and practitioners in pursuing maximum performance from their machine learning systems. A better understanding of these cutting-edge techniques is expected to facilitate the achievement of better and more innovative results in future machine learning applications.

Quantum computing has emerged as a transformative technology with the potential to revolutionize numerical methods in scientific research. This study explores the integration of quantum algorithms to enhance the efficiency and accuracy of computational techniques used in solving complex scientific problems. The objective of this research is to investigate how quantum computing can address limitations in classical numerical methods, particularly in areas such as optimization, simulation, and data analysis. By employing quantum-enhanced algorithms, such as quantum Monte Carlo and quantum machine learning, the study demonstrates significant improvements in processing speed and solution quality. The findings highlight the capability of quantum computing to tackle challenges in high-dimensional computations and provide novel insights into scientific phenomena. These advancements have profound implications for disciplines ranging from physics and chemistry to material science and beyond, paving the way for a new era of computational-driven discoveries.

This research explores various machine learning approaches, including deep learning and ensemble methods, to predict climate change indicators. We focus on temperature and precipitation trends using large datasets spanning multiple decades. By comparing the performance of algorithms like CNN, RNN, and random forests, we identify the most accurate models for specific climate variables. Our findings demonstrate that ensemble models provide better accuracy and reliability, especially for temperature predictions.

This study explores the integration of Artificial Intelligence (AI) into smart home systems using the Android operating system to enhance security, privacy, efficiency, and user comfort. Key security measures include data encryption, robust authentication methods, sandboxing, and AI integration, specifically leveraging Google Assistant for improved privacy controls. Maintenance strategies for smart homes emphasize energy management, device condition monitoring, and enhanced safety features. AI adaptation to user habits enhances productivity and situational awareness, while Android's role in connecting various IoT devices facilitates remote control and energy-efficient recommendations. Methods such as Eco Android, Greensource, byte-code transformations, and automated energy diagnosis tools aid in optimizing energy use. The comparison between smart and non-smart homes highlights the efficiency and convenience of smart homes despite higher installation costs and potential network issues. The development and deployment of an Android-based application, SafeHause, exemplifies practical implementation, emphasizing end-to-end testing, security updates, and user education. The findings affirm that AI integration with Android significantly improves the smart home experience by enhancing energy optimization, data security, and personalized user interaction. Furthermore, the study discusses future trends in smart home technology, such as the potential for more advanced AI algorithms and machine learning techniques to provide even greater personalization and automation. The importance of regular software updates and the role of user feedback in refining smart home systems are also highlighted, ensuring that these technologies continue to evolve and meet user needs effectively.  

This article discusses the integration of business information systems with machine learning technology to enhance decision-making efficiency and accuracy. The objective of this research is to implement machine learning algorithms into business information systems to automatically process and analyze data, predict business trends, and provide more intelligent recommendations. The methods involve collecting historical data from business information systems, applying machine learning algorithms such as regression, classification, or clustering, and integrating the analysis results into the BI platform. The outcome is an enhanced capability of business information systems to optimize supply chains, identify consumer behavior patterns, and optimize marketing strategies based on more accurate predictions. Integrating machine learning technology with business information systems proves to be an effective strategy for improving a company's competitiveness in the current digital era.  

Alzheimer's and Parkinson's diseases are neurodegenerative conditions that affect the brain, with Alzheimer's causing cognitive and behavioral decline, while Parkinson's leads to motor and non-motor impairments. Both diseases have significant impacts on the health and quality of life of patients, with prevalence increasing in recent years. Although the exact causes of these diseases are still unknown, MRI (Magnetic Resonance Imaging) is widely used to detect brain activity and serves as one of the diagnostic methods. With technological advancements, intelligent systems in image processing for image classification have been extensively used and have become a popular field due to their ability to replicate human visual capabilities. Image classification is performed using various supervised learning machine learning algorithms based on the shape, texture, and color of the images. This study employs two Convolutional Neural Network (CNN) architectures, ResNet50 and GoogLeNet, to compare the performance of these models in classifying MRI scans of patients with Alzheimer's and Parkinson's diseases. The results show that the ResNet50 model outperforms the GoogLeNet model, with parameters set to 100 epochs, a batch size of 128, a learning rate of 0.0001, and the Adam optimizer, achieving an accuracy rate of 90%.

The rapid growth of data-intensive applications has posed significant challenges for classical machine learning (ML) algorithms, particularly in terms of computational efficiency and scalability. This study explores the role of quantum computing in optimizing machine learning performance through the implementation of Quantum Machine Learning (QML), specifically using the Quantum Support Vector Machine (QSVM) model. The research adopts a Design Science Research approach, involving problem identification, model development, system implementation, and performance evaluation. Both classical Support Vector Machine (SVM) and QSVM models are developed and tested using benchmark classification datasets. The results indicate that QSVM outperforms the classical SVM model across multiple evaluation metrics, including accuracy, precision, recall, and F1-score. Additionally, QSVM demonstrates improved computational efficiency by reducing training time, particularly when handling high-dimensional data. These improvements are attributed to the ability of quantum computing to utilize quantum kernel methods and map data into higher-dimensional feature spaces, enabling better pattern recognition and classification performance.  Despite these promising outcomes, the study also identifies several limitations related to current quantum hardware, such as noise, decoherence, and limited qubit availability, which may affect scalability and practical implementation. Therefore, further research is required to enhance quantum hardware reliability and develop hybrid quantum-classical models. In conclusion, quantum machine learning offers a promising solution to overcome the limitations of classical approaches, providing enhanced performance and efficiency for complex data processing tasks in future intelligent systems.

This paper presents a comparative analysis of various battery management systems (BMS) in electric vehicles, with a focus on incorporating machine learning techniques to improve battery safety and extend battery life. The study evaluates conventional BMS against machine learning-enhanced models in predicting thermal runaway, state of charge (SOC), and state of health (SOH) under diverse operating conditions. Results indicate that machine learning algorithms outperform conventional methods, providing more accurate SOC and SOH estimations, thus enhancing vehicle safety and longevity.

The aim of this research is to provide a study of the computational thinking process of mathematics learning. This research uses library research methods. The data obtained is the publication of research articles in scientific journals. Data analysis includes three stages: organize, synthesize, and identify The result of previous research is students' computational thinking abilities need to be improved in abstraction and algorithms.      

MSIB (Certified Independent Study and Internship) is one of the activity programs at the Merdeka Campus which aims to help students improve their skills and develop themselves. MSIB appointed Orbit Future Academy as one of the partners in the Independent Study program. Founded in 2016 with the aim of improving the quality of life through innovation, education and skills training. In accordance with its mission, namely "We curate and localize international programs and courses for upskilling, re-skilling youth, and the workforce towards jobs of the future". Partners provide opportunities for students to take Artificial Intelligence programs and study online. Learning consists of eight material courses including Python Programming, AI Technology Logic and Concepts, AI Project Cycle, AI Research Methods, ChatGPT, Professional and Company Ethics, Financial Literacy and ending with a Final Project. The final project scope carried out is the Occupancy Detection in Parking Lot project. This project uses the Computer Vision domain with the selection of the YOLO model in detecting objects and pixel segmentation. The project begins with selecting a dataset using roboflow which then goes through data pre-processing for cloning, annotation and augmentation. Then the model is trained using machine learning and deep learning algorithms to understand patterns and characteristics related to parking spaces. Once trained, the AI model will be validated using test data. This aims to ensure that the model truly recognizes the presence of the vehicle. Next, form the application design in creating an informative interface using wireframes. Then enter the deployment stage so that the system can be accessed widely and easily via the web. Lastly, field testing is to find out the performance of the application that has been designed.      

Computer vision technology is used to improve work safety in the construction industry. The key in this project is the utilization of the YOLO method on the Roboflow platform. In addition to the Convolutional Neural Networks (CNN) algorithm, YOLO efficiently divides the image into a grid and classifies the objects in the grid by bounding box and confidence score. With the integration of YOLO, this project can achieve accurate and fast PPE detection. This project uses the YOLO method to detect head and body parts from input images. The detected body parts are then cropped and processed using the CNN method for classification. This project will also implement computer vision algorithms, including Deep Learning methods that currently have the most significant results in image recognition is CNN method, to automatically detect and monitor the use of PPE. This model achieves mAP 64.1%, Precision 73.2%, and Recall 60.2%. The Streamlit framework was used for deployment, creating a web application for PPE compliance tracking. This project, ''Health and Safety PPE Compliance Tracking'', aims to improve work safety in the construction industry. This project uses Computer Vision technology to detect, monitor, and ensure worker compliance with the use of appropriate PPE. The suggestion is to conduct further trials using other datasets in the form of photos or videos that can be done in real-time by ensuring that the colors of hats and vests do not vary too much to detect the conformity of labeling with PPE use.