Accurate forecasting of carbon prices is crucial for policymakers and market participants, yet remains challenging due to volatility, non-stationarity, and structural breaks in carbon markets. Data preprocessing plays a key role in improving predictive accuracy, but conventional normalization methods such as z-score and min–max rely on static statistics and fail to adapt to evolving dis- tributions. This study introduces Extended Deep Adaptive Input Normalization (EDAIN) into carbon price forecasting. EDAIN dynamically integrates outlier suppression, shifting, scaling, and power transformation, optimized jointly with forecasting models. We evaluate EDAIN with GRU, LSTM, and XGBoost on datasets from KRBN and KCCA obtained via the yfinance API. Experimental results show that EDAIN consistently outperforms static normalization, reducing the RMSE from 2.45 to 0.80, the MAE from 2.09 to 0.63, and the MAPE from 7.07
Social Impact Organisations (SIOs) face persistent challenges in governance and transparency, which impact public trust and operational efficiency. The emergence of Decentralized Autonomous Organizations (DAOs) offers a blockchain-based alternative to traditional governance structures, leveraging transparency, decentralization, smart contracts, and cryptographic security. This paper presents a comparative review of DAO adoption in SIOs, analyzing transparency mechanisms, governance models, adoption barriers, and their impact on the sector. Using a Narrative Literature Review (NLR) approach, the study synthesizes insights from nine selected papers out of 122, focusing on DAO-driven donation tracking, governance models, and regulatory constraints. The findings suggest that DAOs enhance transparency, accountability, and cybersecurity resilience through immutable records and tamper-proof decision-making processes. However, technical complexities, regulatory uncertainty, and financial sustainability remain significant challenges to adoption. Hybrid governance models, which balance decentralization with regulatory compliance, appear promising for overcoming these barriers. The review identifies five key research gaps: (i) empirical validation of DAO models, (ii) decentralized identity and inclusive governance, (iii) integration with AI and IoT, (iv) sustainable financial models, and (v) standardized legal frameworks. Addressing these gaps is crucial for advancing DAO adoption in SIOs, ensuring a scalable and legally compliant governance framework. This study contributes by providing a structured comparative analysis of DAO implementation in social impact initiatives, with emphasis on transparency, accountability, and cybersecurity implications.
In the demanding field of practical and feasible quantum cryptography (QC), a multi-party environment is vital for scalable secure communications. This paper presents a four-party quantum key distribution (QKD) simulation framework based on Optical Frequency Division Multiplexing (OFDM) using OptiSystem™. Building upon our prior two-party QKD model, this enhanced framework utilizes wavelength multiplexers and OptiSystem’s subsystem features to streamline transmitter and receiver designs. The implementation includes a centralized server coordinating signal multiplexing and demultiplexing. We reference real-world fiber-optic QKD studies to evaluate expected key generation rates and quantum bit error rates (QBER). The framework supports future research on multi-party authentication and conference key management, offering a modular and replicable simulation for quantum communication development.
Population aging reduces older adults’ social engagement, leading to loneliness and lower quality of life. Phatic communication—interaction aimed at maintaining social bonds rather than transmitting information—offers a non-pharmacological approach to enhance engagement. This study reviews its concepts, applications, and relevance for older adults, highlighting barriers such as sensory decline, technology complexity, and limited emotional richness. We propose integrating touch perception into everyday communication devices to simulate physical presence, strengthen emotional connection, and compensate for sensory loss. Using literature review, case analysis, and design exploration, findings suggest that tactile-enhanced phatic communication improves intimacy, stimulates engagement, and supports emotional well-being. A conceptual framework linking phatic communication, tactile technology, and social engagement is presented, offering guidance for inclusive, emotionally supportive technology design.
Loop-closure detection is a critical component in LiDAR-based SLAM, as it corrects accumulated drift and improves long-term localization accuracy. While considerable research has focused on designing effective descriptors, the role of clustering and indexing methods in retrieving loop-closure candidates has received little attention. Existing systems primarily rely on KD-Tree, with limited evaluation of alternatives like DBSCAN and Spectral Clustering. This paper addresses this gap through a controlled, head-to-head comparison of KD-Tree, DBSCAN, and Spectral Clustering for organizing and matching LiDAR-Iris descriptors. All methods are evaluated under identical conditions using the same dataset, hardware, and binary descriptor format. Experimental results on a 333-frame dataset show that DBSCAN achieves the fastest total processing time (6.61 s), while Spectral Clustering offers the lowest query time (0.0386 s). In contrast, KD-Tree incurs higher memory usage (+631
The concepts of AI (Artificial Intelligence), AGI (Artificial General Intelligence), Superintelligence, and the Singularity can be analysed in the light of Islamic jurisprudence and Islamic philosophy, where the focus is on the creation, purpose, and limitations of human and non-human entities. The progress in AI technologies raises important questions about creation, purpose, and morality in an Islamic context. Is it possible for machines to mimic the abilities of man’s intellect, wisdom (hikmah), natural disposition (fitrah), and belief system? Mimicking is an imitation game for computers, which need the ability to understand languages, solve problems, make decisions, predict futures, innovate ideas, invent products, and ultimately think, feel, and aspire just like human beings. But the speculation of progress in AI goes far beyond the mimicking capabilities of human beings. The transition from Generative AI to God-like AI and Cosmic AI seems to be reaching toward the power of God. How does Islam conceive this progression in the light of Islamic jurisprudence, philosophy, and Sufism? Its plausibility, and the necessary and sufficient conditions to achieve the stages of this progress, need to be articulated. The door to this discussion must be open to understand the goals and to determine their plausibility in achieving them. Misunderstanding in this issue might lead to shirk , which refers to associating partners with Allah or attributing divine qualities to anything besides Him.
3D printing, particularly Fused Deposition Modeling (FDM), is an advanced manufacturing technology widely used not only in industries like medical, automotive, and manufacturing due to its customization, design flexibility, and faster production but also among hobbyists. However, FDM is prone to errors like layer shifting, leading to material waste and delays. This study develops allow cost smart monitoring system for FDM using a Raspberry Pi 4 to detect and halt faulty prints, reducing waste. The system employs a top-view camera for selective image capture, triggered by layer changes from the printer’s serial output. It compares an actual mask (from edge-detected images) with an ideal mask (from G-Code), calculating pixel deviations. If deviations exceed 2
Learning Management Systems (LMS) have become central to the digital learning experience. But providing timely and reliable technical support remains a challenge for universities with large communities. Students and lecturers frequently encounter technical issues such as login failures, course enrolment errors, broken content links and difficulties with online assessments. All of which place heavy demands on human helpdesks. Traditional solutions such as static FAQ pages or scripted chatbots are often unable to keep pace with the scale and variety of queries. In this context, generative artificial intelligence (AI) offers a promising alternative. This paper presents a comprehensive framework for integrating generative AI into LMS technical support using a Design Science Research (DSR) approach and validated through an institutional case. By combining Retrieval-Augmented Generation (RAG) with conversational models like ChatGPT, the framework delivers context‑aware and source‑grounded answers. It operates seamlessly across multiple platforms, including web widgets and WhatsApp. The architecture is designed with layered guardrails, fallback mechanisms and evaluation loops to ensure accuracy, transparency and trustworthiness. Findings indicate that generative AI chatbots can reduce response times, ease the burden on human staff and improve user satisfaction when deployed responsibly. The contribution is not a one-size-fits-all tool but a structured, adaptable approach that higher education institutions can tailor to their contexts. This study shows how generative AI can be integrated into LMS technical support through a structured framework that reduces response times, eases the burden on helpdesks and improves user satisfaction while ensuring accuracy and trust.
The global incidence of prostate cancer has steadily increased over time, mainly due to the growing proportion of the older population. Early detection of prostate cancer, when it is still confined to the prostate gland, significantly enhances the chances of successful treatment and improves survival rates. This study proposes an automated prostate cancer diagnosis and segmentation system using advanced medical image processing techniques. The most relevant features from preprocessed prostate MRI images are selected using a novel Wild Horse Optimized (WHO) feature selection method to optimise the training and testing performance of the cancer detection system. A suite of deep learning models, including Convolutional Neural Network (CNN), Residual Network (ResNet), and Generative Adversarial Network (GAN), is then utilized to classify MRI images as cancerous or non-cancerous accurately. The final classification result is determined through a voting mechanism that selects the best-performing prediction model. Furthermore, a new technique, Dual Swin Transformer UNet Segmentation (DSTra-UNet), is introduced to precisely segment cancer-affected regions within the prostate images. The proposed system is evaluated using various performance metrics, including accuracy, sensitivity, precision, recall, and error rate, demonstrating its effectiveness and reliability in prostate cancer diagnosis and segmentation.
Melon farming in Malaysia requires careful management of soil moisture and temperature to ensure healthy growth and high yields. Previous studies have shown that surface temperature influences these soil parameters and that real-time monitoring enables smarter irrigation decisions. This study investigates how ambient temperature affects soil moisture and soil temperature, in order to support data-driven irrigation strategies in tropical greenhouse agriculture. The study was conducted over four months in a controlled greenhouse in Terengganu, Malaysia, using six sensor stations. Surface temperature, soil temperature, and soil moisture were recorded every 30 min. After data cleaning and normalisation, quantile regression was used to assess the effect of surface temperature on soil parameters across different moisture levels. The quantile regression model at the median (τ = 0.5) revealed a statistically significant positive relationship between surface temperature and soil moisture (coefficient = 0.25926, p < 0.001, t = 21.51), with a low standard error (0.01205), indicating high model precision. In contrast, the relationship between surface temperature and soil temperature was weak and not statistically significant (coefficient = –0.08152, p = 0.09875), suggesting that soil temperature is buffered from ambient fluctuations. These findings confirm that quantile regression offers a more accurate understanding of soil moisture dynamics compared to traditional linear models. Integrating additional variables such as air humidity and atmospheric pressure could improve predictive irrigation strategies.
Fast chargers play an important role in supporting smart city development by enabling access to mobile applications, location-based services, and real-time data systems. As mobile internet technology advances, smartphones are becoming increasingly powerful, but their energy consumption continues to rise. However, the lack of compatibility between fast-charging protocols from different manufacturers presents a challenge for universal charging solutions. To address this, a multi-protocol fast charging charger has been designed to support a range of fast charging technologies. The charger adopts a quasi-resonant (QR) flyback architecture with a GaN power stage and a dedicated controller. This design allows for higher efficiency and compact size, making it well-suited for modern mobile device demands within the broader context of smart city infrastructure. In this work, an experimental prototype was successfully built. The experimental results show that the fast-charging charger can achieve a fast-charging function with a maximum output power of 30 W and has Type-C and Type-A interfaces. It is convenient for fast charging of various data lines. And support 90Vac 264 Vac wide voltage input in line with the AC input of various countries. The Type-C interface supports up to 20 V/1.5 A output. It meets the limits of U.S. DOE Level VI and EU Regulation 2019/1782 in our measurements. In addition, the compatibility of the charging and discharging fast charging protocol of the system prototype has been tested, and it can be quickly charged with a variety of mobile phones that support the fast-charging protocol. It has advantages of high-power density, high conversion efficiency and strong compatibility.
The World Health Organization’s data indicates a consistent increase in the number of people 60 years of age and over worldwide. According to projections, the number of older individuals worldwide will rise from 1.4 billion in 2030 to 2.1 billion in 2050. Changes in perception and cognition brought on by ageing may affect an aged person’s capacity to identify and convey emotions through facial expressions. There is a dearth of research on the identification of emotions in older persons, dementia or not. As a result, there aren’t many databases available to help with research and experimentation in this field. Older adult’s communication skills may be hampered by their incapacity to communicate or identify their emotions, which could endanger their safety. Therefore, the aim for this research was to develop an application for recognizing emotions in elderly individuals based on facial expressions. To realize the objective, there are three main stages to achieve the objectives including, data collection, facial expression recognition algorithm and decision system. Very few of older people presented in common facial expression for emotion recognition datasets. Thus, in this work, facial expression dataset will be collected using high-definition camera for capturing their faces. The camera will be installed in a room where the monitoring system will monitor their emotions in 24 h. Then, the image is input to the emotion recognition algorithm to recognize the emotion for the day. Finally, if the changes of emotion happen in certain frequent of time, the system will recognize there is an issue with mental health. The subsequent stages involve evaluation and testing the model application in real-time implementation. The expected outcome of this research has the potential to markedly improve the quality of life for seniors living alone. Furthermore, it is expected to provide valuable support to caregivers and healthcare professionals through detailed behavior analyses for the development of personalized care plans.
Fine-tuning large Transformer models for short text classification poses significant computational challenges, limiting their widespread adoption in resource-constrained environments. This paper addresses this issue by providing a systematic empirical evaluation of Low-Rank Adaptation (LoRA) as a parameter-efficient alternative to full fine-tuning. We compared LoRA-augmented BERT, RoBERTa, and DistilBERT against their fully fine-tuned counterparts on three diverse benchmarks: R8, TREC, and SST-2. Our results demonstrate that LoRA dramatically improves efficiency, reducing the number of trainable parameters by over 98
The research proposed a more sustainable production of bioenergy and bio-products from wastewater using microalgae green technology with integration of Energy Informatics and applying computer algorithms through optimisation using Artificial Intelligence (AI) and Machine Learning (ML) to the bioprocess control systems. Energy Informatics in its most simplified definition is concerned with the use of advanced ICT to provide appropriate sustainability solutions. This rather new research area in informatics, however, must be studied with the inclusivity Sustainable Development Goals (SDGs) based on practical situations and must deal with environmental, technological, economic, and societal issues. The importance of energy informatics was highlighted to world community when the European Commission in 2009, recommended the mobilization of in-formation and communication technologies (ICT/Informatics) to facilitate the transition of an energy-efficient, low carbon, neutral emission economy in the world. Thus, this research focused on the investigation of potential optimum treatment method that can be adopted by the water treatment industry with the integration of Energy Informatics for a smart monitoring system towards energy efficiency and cost saving in smart city environment. At the end, the aspiration is a sustainable perspective of microalgal biorefinery, with a techno-economic assessment of the circular economy of microalgae application that can provide an insight into the sustainability, energy and cost-effectiveness throughout the whole process system from wastewater treatment to bioenergy production. The prediction and optimisation (AI, ML) used in the process correlates with influential parameters such as renewable sources availability and the demand of microalgae bio-products that it can create using minimal energy. The study hopes to be able in the near future to create a low electricity through bioelectricity and bioenergy and the development of fuel-cell Power computer chips for small devices such as laptops, tablets and hand phones and control systems in smart homes of Smart Cities.
The successful implementation of the National Address System (NAS) in Malaysia is a critical enabler of inclusive digital transformation, socio-economic development and effective public service delivery. As the country moves toward integrated governance and smart infrastructure (e.g. smart cities and smart villages), a robust NAS forms the foundational layer for location intelligence, last-mile service access, and spatial data interoperability. This paper highlights the dual role of policy frameworks and Technological (AI) enablers in accelerating NAS deployment across diverse geographics in Malaysia-encompassing urban, rural and remote in the five (5) regions studied (Sabah, Sarawak, Northern, East Coast and Central-Southern Region) of Malaysia. Policy enablers such as data standardisation, regulatory mandates, cross-agency governance, and inclusive addressing guidelines, serve as the backbone for institutional alignment and adoption. Parallelly, new technologies (AI and geospatial technologies), are revolutionising address generation, validation and mapping, particularly in areas where traditional addressing is absent or ambiguous. By leveraging on visual-based use-cases – including satellite imagery, drone mapping, and computer vision, this paper demonstrates how AI can intelligently detect dwellings, assigns address coordinates and updates registries in near real-time. The study presents a series of visual narratives based on use-case examples that highlight practical potential integration of AI into NAS workflow – especially for informal settlements such as longhouses and indigenous villages. It also examines the ethical, technical, and social dimensions of AI-led address systems, including issues of privacy and inclusion. The paper concludes with evidence through qualitative data-use cases and expert interviews to include these two (2) foundational pillars: Policy enablers and technological (AI) enablers as NAS implementation success factors in building a connected, inclusive as well as an equitable smart nation and society.
This decade-long review examined BCI-based working memory interventions published between 2014 and 2024, focusing on the design elements that influence usability, engagement, and training effectiveness. While existing studies often report positive outcomes, the design features critical to user experience and long-term applicability remain underreported. Conducted in accordance with the PRISMA 2020 guidelines, this systematic review analyzed 13 eligible studies identified through a structured search across four databases to identify and compare design elements. Continuous visual feedback and personalized neurofeedback thresholds were the most common, with N-back, Sternberg, and Corsi tasks frequently used to modulate cognitive load. Adaptivity was widely implemented, yet older adults were underrepresented, and usability evaluations were rare. Gamification appeared in fewer than half of the studies, often without systematic assessment of its impact. Although many systems incorporated personalization and adaptive difficulty, the lack of user-centered design, especially for older or clinical populations, highlights a gap between experimental prototypes and real-world applications. The review concludes with preliminary design element guidelines to inform the development of effective, accessible, and engaging BCI-based working memory systems.
This study explores the transformative nature of Internet of Things (IoT) technologies in the smart city development process in China by using qualitative analysis of the opinions of the people working in related industries. Using an exploratory research design, the study encompassed a semi-structured interview with the professional opinion of seven respondents, comprising contractors, building architects, and quality control managers, who worked on smart city projects. The study investigates the practical applications and the strategic implementation of IoT in urban environments by conducting qualitative interviews with professionals in the industry, such as contractors, architects, and quality control managers. The results show that IoT plays a pivotal role in enhancing the efficiency of workflow, controlling resources, and the rule-based governance based on data. Nonetheless, issues such as infrastructure integration, financial limitations, and data protection remain threats to optimal deployment. Comparing the findings of the study with the existing literature helps in the identification of similarities and discrepancies in existing literature, which provides a deeper context of the current need for practice as well as for the future.
This study investigates how blockchain technology can be leveraged to address challenges associated with digital transformation, with the aim of enhancing knowledge sharing in higher education. It focuses on six core features of blockchain—transparency, security, decentralization, immutability, traceability, and smart contracts—and outlines their potential to support trust-building, streamline verification processes, and improve incentive alignment. The study positions blockchain as a trusted knowledge infrastructure that offers not only functional improvements but also a structural framework for rethinking how institutions coordinate and govern knowledge. The paper highlights blockchain’s potential to provide long-term structural solutions to persistent problems in knowledge governance, offering both theoretical foundations and practical implications for advancing digital systems in higher education.
Recognizing emotions in children presents a unique challenge, as their emotional expressions are often subtle, variable, and influenced by developmental factors. Traditional recognition methods frequently fall short in capturing these complexities. In response, this study introduces a comprehensive deep learning-based approach designed specifically for classifying children’s emotions. The methodology includes systematic data collection, advanced feature extraction, and robust classification processes. A curated dataset was developed featuring three primary emotional states, happiness, sadness, and anger. Deep neural networks were employed to automatically identify and learn meaningful patterns from the data, allowing for the extraction of high-level features essential for accurate emotion recognition. To ensure practical applicability, the trained model was integrated into a web-based interface, enabling real-time emotion analysis through API interactions with user-submitted images. Evaluation results indicate that the model achieves reliable predictive performance. This research not only contributes to improving the accuracy of automated emotion detection in children but also provides a foundation for future innovations in educational technology, mental health monitoring, and affective computing tailored to young users. Experiments conducted on real-world datasets demonstrate that the proposed system outperforms conventional approaches in accuracy with 95
Virtual teaching tools for vocabulary learning have been implemented in classrooms to comply with new standards. However, empirical evidence is still limited regarding their effectiveness for ESL primary students and learners’ perceptions of their use. This study examines the benefits of employing such tools for vocabulary development among English as a second language (ESL) primary students and investigates students’ attitudes towards these materials. The goals are to evaluate how virtual teaching tools affected ESL primary students’ vocabulary development and how they felt about using them. A descriptive research design was used, with pre- and post-test scores analysed using descriptive statistics. Score comparisons were made by percentage and total improvement scores, while questionnaire mean scores and standard deviation were analysed descriptively. Qualitative data from semi-structured interviews were transcribed, categorised, and coded using content analysis. Students’ mean pre-test score was M = 18.20; SD = 1.42, and their post-test score was M = 43.63; SD = 2.06, showing a mean improvement of 25.43 and a total improvement of 139.73
