Journal of Advanced Research Design

A Real-Time Hospital Selection Framework for Multi-Chronic Disease Patients in Telemedicine

Noor Farhan Abbas — 2025-08-13

The development of an integrated healthcare system has become a personally important issue in the healthcare industry because of the rapid increase in the prevalence of several chronic diseases. The provisioning of personalized healthcare service (PHS) is one of the challenges facing researchers in the fields of telemedicine and the Internet of Medical Things (IoMT) and is related to several issues. One of the main issues is providing the remote patients with the required pre-hospital services which includes the determination of the patient’s emergency level and selecting the proper hospital. This study aims to improve the provisioning of personalized healthcare service for the multi chronic disease elderly patients who use telemedicine system through utilizing medical sensors and non-sensors IoMT devices, That includes determination of the patient’s emergency level (triage level), identifying the disease, and selecting the proper hospital in terms of availability and location. The proposed framework called Multi Sources Healthcare Architecture-2 (MSHA-2). The proposed framework accommodated the patient with different contingency levels based on seven medical devices, namely, sensors for blood glucose, blood pressure, ECG, SPO2, body temperature and respiratory rate and input medical texts. The framework proposes two computational algorithms to achieve the required tasks. The simulation of the framework is proposed in Baghdad, the capital city of Iraq, as the real physical locations for 34 private and governmental hospitals are determined using Google map coordinates. The appropriate hospital for the patient was identified in accordance with the patient’s coordinates, which were taken from a GPS sensor. According to common evaluation performance criteria, the proposed framework outperformed two benchmark models and obtained 90% of the total services, whereas the benchmark studies obtained 45% and 45% respectively. This study proposes a unique advance healthcare solution which supports the healthcare digital transformation plans in Iraq. Selecting the proper hospital to the patient can save patients’ lives. Future work will integrate patient medical records and improve hospital selection during disasters to enhance real-time healthcare services

Integrating facenet and extreme learning machines for enhanced face identification: addressing real-world challenges with advanced image processing techniques

Areej A. i Abed — 2025-08-13

Face identification forms an important area of computer vision due to its application in many real-time applications related to the security and surveillance of law enforcement agencies. Deep learning algorithms have revolutionized building face identification systems with high accuracy and speed. There is a dire need for systems that achieve more reliability and can handle a diverse range of scenarios. This work uses a junction of FaceNet, a convolutional neural network (CNN)-based model for feature extraction, and extreme learning machines (ELM) to form a face identification system. The projected performance of the proposed system is expected to outperform that of the existing systems in terms of accuracy and resilience. The improved performance of the proposed systems is rooted in FaceNet and ELM, which can capture intricate facial features and patterns accurately. The ELM runs very fast with a single hidden layer feed forward neural network. Extensive experiments were conducted on the Youtube-faces dataset, and images were captured in real time. The proposed method had a recognition accuracy of 99.1 percent, a precision of 98.5 percent, a recall of 97.8 percent, and an F1-score of 98.1 percent. Further, we have also applied pruning and quantization to compress the FaceNet+ELM model for its efficient performance on low computational power devices. Pruning reduces redundant weights and neurons, while quantization converts parameters from 32-bit to 8-bit, greatly reducing the model size and increasing the inference speed.

Design of a 100 Watt High Gain Cascaded Modified Boost Converter with Continuous Input Current

Mohd Nadzri Mamat — 2026-02-15

The conventional boost topology can be constructed to achieve high voltage gain by employing an extreme duty cycle, which can lead to inefficiencies, heightened voltage stresses, and instability. In order to address these challenges, this study introduces a high gain cascaded modified boost converter. This converter is specifically tailored for low voltage input sources necessitating high voltage amplification, such as those found in renewable energy systems. The design offers advantages including simplified switching control, high voltage gain, consistency across a broad duty cycle range, continuous input current, and scalability for achieving further gains. The paper elaborates on the operational principles, design considerations, and specifications of this proposed converter. Validation through Simulink simulations has demonstrated that the proposed design can effectively operate within the specified parameters, delivering a well-regulated output of 96 V and 1.04 A from a 12 V input source, with minimal output ripple of less than 0.4%.

Sustainable Leaf Plant Disease Based on Salp Swarm Algorithm for Feature Selection

Hamsa E. d Mahmoo — 2025-08-13

Sustainable plant protection and the economy of plant crops worldwide depend heavily on the health of agriculture. In the modern world, one of the main factors influencing economic growth is the quality of agricultural produce. The need for future crop protection and production is growing as disease-affected plants have caused considerable agricultural losses in several crop categories. The crop yield must be increased while preserving food quality and security and having the most negligible negative environmental impact. To overcome these obstacles, early discovery of satisfactory plants is critical. The use of Advances in Intelligent Systems and information computer science effectively helps find more efficient and low-cost solutions. This paper proposed a multiclass classification model that aims to detect diseases in three types of fruit using the leaves plant images dataset. These three types of fruit are (Apple, Cherry, and Strawberry) where Apples have three disease dataset categories (Apple Scab, Black Rot, and Cedar Rust) as well as healthy apple dataset, Cherry have Powdery Mildew disease dataset category and healthy dataset, and Strawberry have leaf Scorch disease dataset category and healthy dataset. These datasets are based on the Kaggle website. These multiclass classifications need several steps of processing; the first step is preprocessing the dataset by resizing all images to the same size, segmentation, and removing noise; then, feature extraction from color and texture features; the next step is feature selection to find optimal features by using the Salp Swarm algorithm (SSA); and classification by using machine learning models (Random Forest), (CatBoost), and (XGBoost). In the final step, evaluation of the performance was used to select several matrices: Accuracy, precision, recall, and F1-score.

Deep Learning Approaches for Accurate Diabetic Retinopathy Detection and Classification: Comparison Study

Mohammed Ibrahim Mahdi — 2025-08-13

In order to prevent irreversible blindless among adults aged 18-65, it is imperative to accurately diagnose and treat diabetic retinopathy (DR) as early as possible. As such, the present study endeavoured to compare the efficacy of four deep learning (DL) models; namely, convolutional neural networks (CNN), residual networks (ResNet), inception architecture (IA), and densely connected convolutional networks (DenseNet); at detecting and classifying DR into different levels of disease severity. The Kaggle DR Detection dataset was used to assess the classification accuracies while a loss function (LF), that combines the loss of cross-entropy with additional penalties for classification errors, was introduced to overcome class imbalance issues and improve the performance of the four examined DL models. The DenseNet model had the highest accuracy, recall, precision, F1-score, and area under the receiver operating characteristic curve (AUC-ROC) of the examined models, by scoring 90, 89, 88, 88%, and 0.92, respectively. This performance was closely followed by that of the ResNet model. The findings indicate that the architecture of the model, especially that of models that will be used for medical image classification (MIC), must be taken into account when selecting which model to use. Furthermore, the proposed customised LFs enhanced the precision and resilience of the examined DL models. Screening tools that can accurately diagnose DR early, with limited to no intervention from an ophthalmologist, will enable them to treat patiently significantly earlier, thus improving patient outcomes. As such, the development of such models is imperative. However, various datasets should be used to substantiate the accuracy of these models. Their efficacy could also be improved by combining supplemental clinical data as well as examining the use of hybrid architectures.

Adaptive Hybrid Deep Learning Model for Real-Time Anomaly Detection in IoT Networks

Esraa Saleh Alomari — 2025-08-13

The exponential growth of the internet of things (IoT) networks has posed great challenges in maintaining security because of the diverse and dynamic nature of connected devices. These complex environments often defeat traditional anomaly detection methods. In this paper, we present an adaptive hybrid deep learning framework incorporating Variational Autoencoders (VAEs) and Deep Neural Networks (DNNs) enhanced by XGBoost for feature selection for real-time anomaly detection and classification in IoT networks. The VAE component efficiently reduces high dimensional input data into a lower-dimensional latent space that maintains essential network traffic features while managing information loss for storage optimization and improved accuracy in detecting anomalies. Afterward, XGBoost is utilized to choose top 10 significant features with respect to feature selection. The DNN portion uses these latent features to detect as well as classify anomalies; it is therefore trained on how to identify complex patterns within the selected latent features so as to effectively tell apart malicious activities from normal network behaviors. Test Accuracy of 92.8% was achieved by the proposed VAE-DNN model, which displays its efficiency in IoT real-time anomaly detection. Mode also helps improve model accuracy when coupled with XGBoost based feature selection techniques

Advancing CO2 Separation through Polyphenylene Sulphide-Based Fillers: A Preliminary Review

Afdhal Junaidi — 2025-07-21

Efficient removal of Carbon dioxide from biogas, which is consist of methane (CH4) and CO2, is essential to achieve the biomethane standard of more than or similar 95% CH4. Conventional separation methods, such as cryogenic distillation, pressure swing adsorption and amine-based absorption, face significant challenges in terms of energy consumption, complexity and scalability. Membrane-based separation has emerged as a promising alternative, with polymeric membranes offering specific advantages for CO2 removal. However, traditional polymeric membranes encounter issues such as limited selectivity and compatibility, often requiring modifications or the incorporation of fillers. This review explores the innovative application of polyphenylene sulphide (PPS) including porous carbon materials derived from PPS (PCs) and nitrogen-sulphur co-doped (NSPCs) modifications, as organic fillers in polymers that focus on polysulfone membranes to enhance CO2 separation. This review presents state-of-the-art advancements in biogas separation membranes. PPS-based fillers offer the potential to overcome limitations associated with traditional inorganic fillers, providing improved separation efficiency, compatibility and sustainability. This review also discusses the broader implications of enhanced biogas utilization, emphasizing its relevance to sustainable energy goals and environmental policy. The integration of PPS-based fillers addresses limitations associated with conventional fillers, offering enhanced compatibility, increased separation efficiency and improved scalability. By supporting sustainable energy objectives and aligning with environmental policies, this approach offers a pathway toward next-generation membrane technologies for efficient, scalable and sustainable biogas purification.

Design, Fabrication and Performance Test of Arduino-Based Automatic Cutting Tool for Coconut Shell Charcoal Briquettes

Samsudin Anis — 2025-07-21

This study aims to design and fabricate automatic briquette-cutting equipment that uses Arduino technology. This is necessary to ensure that the briquette-cutting process is more efficient and rapid, with accurate cutting results and a uniform dimension. The research processes include design criteria, morphological matrix, product concept, decision-making matrix, geometric design selection, fabrication of the selected design and performance testing. According to the decision-making matrix, concept 3 emerges as the most appropriate concept because of its numerous advantages, such as its robust structure, ability to endure vibration, dependable performance, readily accessible components and superior precision in cutting briquettes. The study's findings demonstrate that the autonomous cutting tool has exceptional performance. Using the proximity sensor's input, the Arduino can precisely command the servo motor to cut the briquettes using the cutting plate. The initial performance test yielded a cutting consistency of 68%. Nevertheless, the second performance test yielded a cutting consistency of 78%.

An Advanced Research Based on Machine Learning Techniques with Variant ARIMA Methods for Identifying Most Severity COVID-19 Data

You-Shyang Chen — 2025-07-21

The COVID-19 epidemic has been a hot topic for a time-series forecasting in the last three years (2020-2022) data, and their national conditions and environmental backgrounds of different countries have different; thus, it is difficult but necessary to find a more suitable time-series model for prediction of relevant data within a short period of certain time. Therefore, tailored to countries to find out such an intelligent model to be able to predict the confirmation cases of coronavirus for different countries has an interesting and important challenge. COVID-19 studies have used traditional statistics-based research architectures, and there are several core problems: (1) it is lack of effectively organizing time-series forecasting models with machine learning techniques for COVID-19 data; (2) it may require different forecasting models depended on characteristics of different yearly COVID-19 data; (3) it is a valuable issue to find out a well-off forecasting model for modulating different parameters of COVID-19 data contexts. This study is trigged by such a research motivation to propose an integrated approach of different three time series-based models, such as autoregressive (AR), moving average (MA), and autoregressive integrated moving average (ARIMA), along with 12 machine learning models for addressing COVID-19 data applications. The study collects the data of confirmation cases from 2020-2022 and divides it into different sub-datasets from the most severities COVID-19 of countries to assess the effect of the proposed model. The evaluation standard is an effective indicator of mean absolute percentage error (MAPE). For the empirical results, it is found that different predictive techniques for the same country’s epidemic data have different prediction error rates; it is proved that different COVID-19 data of different countries needs different forecasting methods. The empirical results provide valuable information as a useful reference for different interest considerations to the interested parties. It is a good issue to differentiate from the existing works, the study highlights an organization method for three time-series models and 12 machine learning techniques in matching a COVID-19 data application; thus, this study has a clear contribution of a technical and applicable innovation on benefiting COVID-19 data fields.

Multi-Criteria Assessment for Hydrogen-Based Decarbonisation Towards Net-Zero Emission for Eco-Industrial Parks in Malaysia

Mohamad Azreen Firdaus Abd Aziz — 2025-07-21

This study employs the Analytic Hierarchy Process (AHP) methodology to conduct multi-criteria decision-making and determine the hydrogen-based energy transition model for the Eco-Industrial Park's decarbonization, based on the Malaysian industrial landscape. This research study is performed by incorporating the integration criteria of the industry supply chain and enabling parameters of funding, infrastructure, regulation, skills, and technology in the computational process of the AHP, and it is ranked accordingly. Two aspects are being considered for the hydrogen energy-based transition: the fuel switching option from the existing energy supply source at the industrial park comprising electricity and thermal, and the sustainable method for the hydrogen production source. The top three AHP results for the electricity and thermal energy indicated that the National Grid is ranked the highest at 0.87, followed by natural gas at 0.82 and biomass at 0.74 for the fuel switching into hydrogen for the energy transition. Meanwhile, the top three results for the hydrogen supply source indicated that the industrial park's best option for hydrogen production is the Green Hydrogen via electrolysis process from the Large-Scale Solar at 0.94. It is followed by Grey Hydrogen via Steam Methane Reforming from the natural gas source at 0.82 and Orange Hydrogen via biomass gasification at 0.82. The overall ranking process for the energy supply system at the industrial park provides a systematic priority and basis for the fuel switching strategy of the electricity and thermal energy and the best selection for hydrogen production towards the carbon emission reduction at the industrial park level. This method can assist the decision-makers in sustainability energy planning as part of the energy transition to transform the industrial park into an Eco-Industrial Park.

Development of Interactive Multimedia Learning Software (i-Tutor Bahasa Melayu) for Rural Secondary Schools in Malaysia: A Needs Analysis

Kenneth Robin — 2025-07-21

The Malay language is crucial in Malaysia as the national language, language of knowledge, and language of unity. This study investigates the need for developing interactive multimedia learning software for teaching Malay grammar in rural secondary schools. The objective is to address the challenges faced in learning Malay grammar, the use of technology in education, and the existing digitalization gaps in rural areas. Methodology involved surveys and interviews with Malay language teachers to assess students' interest, teaching methods, challenges, software availability, and essential components for multimedia software development. Findings reveal moderate student interest, a blend of traditional and emerging teaching methods, significant challenges in mastering grammar, limited learning software, and consensus on necessary multimedia components. Conclusions highlight the potential for innovative teaching methods, the importance of overcoming learning challenges, and the need for better technology integration in rural education. This study emphasizes the digital divide between urban and rural schools and proposes tailored technology solutions to enhance Malay grammar education in rural areas.

Digitalization Application in Crowdfunding: A Systematic Review

Nurul Aien Abd Aziz — 2025-07-21

The digitalization of crowdfunding has reshaped the fundraising landscape, revolutionizing how projects and ventures secure financial support. This transformation has brought forward innovative platforms, altered user engagement and redefined the dynamics of financial inclusion. Digitalization in crowdfunding has expedited procedures and increased fundraising efforts that change the financial landscape. The purpose of this article is to provide a comprehensive review of the current body of knowledge about the digitalization of crowdfunding. The PRISMA approach was used to analyse an extensive compilation of empirical papers regarding the digitalization of crowd fundraising. These articles were obtained from the Scopus and Web of Science (WoS) databases using a search string of relevant keywords. A screening process was conducted to assess the eligibility of these articles, resulting in a total of 30 articles that were deemed fit for further analysis. The findings revealed three key themes concerning digitalization in crowdfunding research: (1) crowdfunding models and adoption, (2) technology and entrepreneurial financing and (3) social and cultural influences on crowdfunding. This study sheds light on the complicated interaction between blockchain, fintech and crowdfunding by investigating the dramatic implications of digitalization on crowdfunding dynamics. Crowdfunding also has prominent impacts on venture capital investments, China's dynamic digital finance ecosystem and Malaysian public schools' investment intents. Additionally, factors like trust, social effect and effort expectation are highlighted in the study, thus informing future crowdfunding techniques and platform operations in advancing financial technology and crowdfunding.

Identification of High-Risk Factors and Advanced Detection of Diabetes Utilizing a Hybrid Conv-LSTM Model

Alhuseen Omar Alsayed — 2025-07-21

Diabetes is a chronic disease that causes various damages to the human body, making early detection crucial. Hence, to address this issue, the current study utilizes hybrid convolutional long-short term memory (Conv-LSTM) Network which help to detect and classify diabetes at the early stages. The proposed Conv-LSTM enhances the model’s prediction by allowing CNN for spatial extraction of feature and LSTM for temporal extraction of feature from the input data. The proposed approach is applied to BRFSS dataset through the implementation of a computerized system for early identification of diabetes. The data gathered from the BRFSS dataset undergoes pre-processing step to ensure that it is suitable for further processing. The pre-processed data is then fed into the Conv-LSTM model which is trained to identify diabetes based on the risk factor. The efficacy of the proposed CGRU framework has been proven by validating the experimental findings with the existing state-of-the-art approaches. Compared to existing methods like machine learning, the proposed framework exhibited better performance. This demonstrates the efficacy of the Conv-LSTM architecture for diabetes prediction achieving high accuracy rate of 98.5%. The approach successfully identifies people who are at high risk of acquiring diabetes and achieves high accuracy in early diabetes detection, allowing for prompt intervention and individualized healthcare treatment.

Development of Digital Games as Problem Solving-based Learning Media using Game Development Life Cycle

Pratama Benny Herlandy — 2025-07-21

The development of digital games as learning tools holds significant potential for enhancing student motivation and engagement in the teaching and learning process. As problem-solving-based learning media, digital educational games offer benefits such as fostering critical thinking, creativity, and problem-solving skills. Despite these advantages, higher education at Muhammadiyah University of Riau remains predominantly conventional, with game-based learning methods yet to be implemented. This study aimed to develop a digital educational game that supports constructive student learning. The research utilized a development research approach, employing the Game Development Life Cycle (GDLC) method. The findings demonstrated the GDLC method's effectiveness in creating educational games for students. Alpha testing revealed that all game features functioned properly, and usability testing, conducted with 50 students using the User Experience Questionnaire (UEQ), yielded positive results. The average scores were as follows: Attraction (5.94), Clarity (6.02), Efficiency (6.38), Accuracy (6.12), Stimulation (5.66), and Novelty (5.80). Implementation trials employed a quasi-experimental design with control and experimental classes to assess the impact of the developed educational game on students’ problem-solving abilities. The results showed statistically significant differences between the experimental class, which used the game, and the control class, which did not. The t-count value for the experimental class was 55.934, compared to 43.987 for the control class, indicating a stronger improvement in problem-solving skills among students who engaged with the educational game. This study highlights the potential of digital educational games as effective tools for modernizing higher education and improving student outcomes.

Predicting Azimuth and Tilt Angles of Photovoltaic Panels with Random Forest and Support Vector Machine: A Study in the Shore Area of South Papua, Indonesia

Yohanes Letsoin — 2025-07-21

Indonesia has various potential renewable energy sources such as solar energy, which can generate more than 200 GW. However, this promising energy source tends to be utilized in small numbers, i.e., less than 100 MW, compared to the potential energy. The solar photovoltaic (PV) systems significantly depend on the location, such as the optimal configuration of azimuth and tilt angles, which determine the amount of solar irradiance captured by the panels. Even though numerous studies and practical applications of PV systems have been investigated globally, unfortunately, there is a scarcity of research focusing on PV technology, specifically in the context of South Papua, where its usage remains unexplored, for example, in the optimal configuration of azimuth and tilt angles. Therefore, this research focuses on optimizing the tilt angle and azimuth of PV panels located in coastal areas of South Papua by utilizing PVSyst software and annual radiation estimation. Moreover, the elevated solar radiation levels in the area establish it as a key site for installing PV systems; on the other hand, the distinctive geographical and weather conditions require accurate positioning of panels to achieve the best possible efficiency of PV systems. To address this, we employed Random Forest (RF) and Support Vector Machine (SVM) models in a Python environment to analyze the impact of azimuth and tilt angles on the global collection plane of solar panels. The dataset included azimuth angles ranging from 0° to 60° and tilt angles from 8° to 45° representing a comprehensive range of possible configurations. Our findings indicate that an azimuth angle of 60° and a tilt angle of 8° yield the highest global collection plane values, making it the most effective configuration for maximizing solar energy capture in this region during the dry season. Conversely, in the wet season, the optimal tilt angle shifts to 30° with an azimuth angle of 0°. Furthermore, the Random Forest (RF) model outperformed Support Vector Machine (SVM) in predictive accuracy, as evidenced by a lower mean squared error. This study not only provides valuable insight into the optimization of solar field orientation but also highlights machine learning techniques in improving PV system performance. By optimizing solar energy capture, this research supports Indonesia’s broader goals of transitioning to renewable energy sources, contributing to sustainable development and energy security in South Papua, Indonesia.

Infectious Disease Risk Assessment using Different Enhanced-FMEA Approaches

Mohd. Zulhilmi Firdaus Rosli — 2025-07-21

Three years before, the global spread of COVID-19, originating in China, rapidly impacted numerous countries, causing a surge in cases and fatalities. Governments worldwide encountered significant challenges not only in healthcare but also across various sectors. As an alternative to reducing the spread of COVID-19, many researchers implemented the Failure Mode and Effect Analysis (FMEA) method to mitigate the associated transmission risks in specific settings. However, this method did not thoroughly examine the process of assigning importance weights and expert judgments to the risk factors, potentially limiting the comprehensive outcome of the risk assessment. This paper discusses the comparison between FMEA, fuzzy-based FMEA, and FMEA-based fuzzy TOPSIS to assess their effectiveness in handling infectious diseases. The longhouse at Pasai Siong, Sarawak, was chosen as a case study due to being one of the most significant clusters during the pandemic in Sarawak. The study's findings suggest that all risk assessment methods unanimously identify the living room (F 3.1) as the most critical area with the highest transmission potential, emphasizing the necessity of prioritizing this area. However, slight variations in rankings across methods were observed due to the distinct approaches taken by each assessment method.