UTAR Institutional Repository (Universiti Tunku Abdul Rahman)
Not a member yet
6132 research outputs found
Sort by
Parking reservation and management system for UTAR Kampar campus
Get PDFThis project is dedicated to developing a smart parking system at UTAR Kampar campus. The
difficulty in finding an optimal parking space has always been a problem for campus users who
drive. The increasing demand for parking spaces has resulted in a waste of time and traffic
congestion, especially during peak hours. Therefore, research is conducted in this report to
examine existing parking systems, parking space availability detection methods,
recommendation system, and reservation strategies. Based on the literature review, the
proposed solution is to develop a mobile-based parking reservation and management system to
address traffic congestion, reduce search time and solve parking inefficiency. The system will
give recommendations to users, enable them to reserve parking using credit, view real-time
availability of spaces, and receive GPS navigation to their reserved spots. The proposed system
is developed using Flutter for the user interface with integration of Firebase and the Google
Maps service into a mobile application to demonstrate how smart reservations can reduce
search times, optimize parking resources in campus and improve user experience
Sustainable food consumption: A perspective from Generation Z
Get PDFThe study involves the examination of consuming sustainable food from the perspective of Gen Z in Malaysia, aiming to promote the adoption of sustainable food practices. The study utilizes the TCV, which encompasses 5 values. The data were gathered via questionnaire from 305 respondents. The data were analyzed using Smart PLS 4.1.1.4. From the findings, a substantial relationship occurred between functional, epistemic, and conditional value with sustainable food consumption. While the social and emotional value has been found to have a weak correlation with sustainable food consumption. The study further provides insights and understanding into encouraging sustainable food practices among young consumers in Malaysia. Keywords: Theory of consumption value (TCV); sustainable food consumption; Generation-Z; Partial Least Squares Structural Equation Modeling (PLS-SEM); sustainability Subject Area: HF5410-5417.5 Marketing. Distribution of Produc
Development of glass waste and glove former–based geopolymers for carbon emission reduction
Get PDFExploration of geopolymers as substitutes for regular Portland cement (OPC), which emits a lot of carbon dioxide (CO₂), has been spurred by the growing need for environmentally friendly building materials. This study investigates the feasibility of utilizing glove former waste (GFW) and glass waste (GW) as aluminosilicate precursors in geopolymer production, with the aim of reducing cement dependency and mitigating environmental impacts. Three mix designs were prepared: 100% GFW, 100% GW, and a blended mix containing 50% GFW and 50% GW. The specimens were cured under ambient conditions for 7, 14, and 28 days, with an additional 24 hours of oven curing at 100 °C after 28 days to assess strength enhancement. Compressive strength tests revealed limited early-age performance (0.2–0.4 MPa), but strength improved at 28 days, with the blended mix achieving the highest value of 1.32 MPa, followed by GFW (1.10 MPa) and GW (0.95 MPa). The oven curing process further enhanced strength development, confirming the positive influence of thermal activation. Chemical and environmental analysis indicated that GFW and GW are rich in silica and alumina, making them suitable precursors for geopolymerization, while also offering a pathway to waste valorization and carbon footprint reduction. The findings suggest that GFW and GW can serve as viable alternative raw materials in geopolymer binder systems, contributing to sustainable construction practices and climate change mitigation. Keyword: geopolymer, glass waste, glover former waste, ordinary portland cement, carbon emission Subject area: TA401-492 Materials of engineering and construction. Mechanics of material
Anomaly Detection in Surveillance Videos
Get PDFIn the present society, video surveillance systems are rapidly evolving with intelligent video analytics to improve public safety. With the increasing installation of surveillance cameras in both public and private spaces, there is a growing reliance on continuous monitoring to ensure public safety. However, human-based monitoring is labour-intensive and inefficient. Video anomaly detection (VAD) plays a vital role in modern surveillance systems by
automatically identifying unusual events in video streams. This study focuses on developing a lightweight and efficient VAD framework that supports both binary and multiclass detection. The proposed system, AnomLite combines MobileNetV2, a lightweight Convolutional Neural Network (CNN) for spatial feature extraction, and Long Short-Term Memory (LSTM) for temporal modelling. By leveraging the strengths of MobileNetV2 in extracting efficient spatial features and LSTM in capturing temporal dependencies in video sequences, the model detects anomalous events across various classes. The
system trains on two datasets: UCF-Crime, which contains real-world CCTV footage, and XD-Violence, which includes video content from movies and YouTube. Preprocessing steps are employed to ensure the model performs well under varying data conditions. The evaluation of the proposed model shows
strong performance on the first dataset, achieving an ROC AUC of 0.99 and an average precision of 0.99 on UCF-Crime. The model demonstrates strong performance on another well-known dataset in video anomaly detection, achieving an ROC AUC of 0.98 and an average precision of 0.97 on XD Violence. The model also achieves high accuracy of 94% on UCF-Crime and
93% on XD-Violence, with strong F1 scores across both datasets (F1-Micro 0.93 on UCF-Crime, 0.89 on XD-Violence). The model achieves high per-class accuracy across the UCF-Crime dataset, with 10 out of 14 classes exceeding 0.95 accuracy and several classes, such as Arson, Explosion, Fighting, Shooting, and Vandalism, reaching a perfect accuracy of 1.00, demonstrating the model’s strong and consistent performance in detecting diverse types of anomalies. Moreover, the model performs well on the XD-Violence dataset, with accuracies ranging from 0.79 to 0.95. It shows highest accuracy on Car Accidents (0.95) and strong performance across other classes like Abuse, Riot,
and Fighting, indicating its effectiveness in handling diverse anomalies. Additionally, the model is optimized for inference through quantization. With a reduction of around 70% in model size through model compression techniques
such as quantization, the flexibility of the model is further improved, particularly for low-end devices. These results highlight how deep learning techniques, such as SMOTE, data augmentation, and advanced loss functions like cross-entropy loss, contribute to high accuracy and effective performance
in automating surveillance tasks, even when dealing with highly imbalanced datasets. Data augmentation techniques that simulate real-world conditions enhance the efficiency of anomaly detection systems in practical applications.
Keywords: Video anomaly detection, deep learning, edge computing, artificial intelligence, neural network
Subject Area: TK7885-7895 Computer engineering. Computer hardwar
The behavioural patterns of Malaysians towards the adoption and satisfaction with digital-only banks
Get PDFThe rapid evolution of financial technology has brought about a profound transformation in the banking sector, with digital-only banks emerging as a key force reshaping the industry. Operating exclusively online without physical branches, these institutions are redefining how banking services are delivered and accessed. This transformation is not merely driven by technological advancements but also reflects a shift in consumer preferences toward more convenient and personalized financial solutions (WEI YET, 2024). In Malaysia, the rise of digital-only banks has gained significant momentum, supported by growing consumer demand and proactive regulatory efforts.
Digital-only banks have existed for over two decades, with their establishment beginning in the late 1990s and early 2000s in countries like the United States, Europe, and Japan. In Asia, digital-only banks gained traction during the 2010s and 2020s, notably in China, South Korea, and Singapore (Yoon & Lim, 2020). Malaysia has also joined this trend due to advancements in its digital economy and financial technology, though the concept is still in its infancy within the country. As of April 2022, Bank Negara Malaysia (BNM) has issued licenses to five digital-only banks. However, before these banks can begin operations, they must undergo a readiness validation and audit by BNM within 12 to 24 months of the license announcement (Five Successful Applicants for the Digital Bank Licences, 2022).
To promote financial inclusion, BNM envisions these digital-only banks offering affordable and accessible financial services to underserved segments, such as small businesses, low-income households, minorities, gig workers, and youth. Additionally, these banks are expected to reduce transaction costs, provide essential financial services, and enhance digital literacy among consumers. By doing so, they aim to make banking more inclusive, improve service quality, boost employment, and reduce poverty (Abdul-Rahim et al., 2022).
The rise of digital-only banks in Malaysia represents a significant transformation in the banking sector, offering innovative, cost-effective, and convenient financial solutions. However, the adoption of digital-only banks remains inconsistent among Malaysians, influenced by various behavioural, cultural, and systemic challenges. While younger, tech-savvy generations readily embrace these platforms, older demographics and rural populations exhibit hesitance due to limited technological literacy, a preference for traditional face-to-face banking, and concerns about cybersecurity (Dharamshi, 2018).
One notable behavioural trend among Malaysians is their high sensitivity to trust and security when adopting financial technologies (Dharamshi, 2018). Cybersecurity threats, including phishing, mobile malware, and password hacking, heighten scepticism toward digital-only banks. According to PWC’s Global Economic Crime and Fraud Survey 2018, 14% of respondents globally who identified cybercrime as the most disruptive fraud reported losses exceeding 100 million. In Malaysia, similar concerns persist, with users expressing anxiety about the safety of financial transactions on fully digital platforms (Price Waterhouse Coopers, 2018). These fears are compounded by incidents involving sophisticated malware that mimic legitimate apps to steal sensitive information, deterring widespread adoption
Green synthesis and characterization of copper oxide (CuO) nanoparticles using barks of soursop (Annona muricata)
Get PDFThis study presents the green synthesis of copper oxide nanoparticles (CuO NPs) utilizing the bark extract of Annona muricata as a natural reducing and stabilizing agent, with copper(II) nitrate trihydrate serving as precursor salt and varying calcination temperatures (300, 400, and 500°C). The effect of different calcination temperatures on the structural, morphological, and optical properties of the CuO NPs (CuO-300, CuO-400, and CuO-500 NPs) was characterized using Ultraviolet-Visible Spectroscopy (UV-Vis), Fourier Transform-Infrared Spectroscopy (FT-IR), Field Emission Scanning Electron Microscopy (FESEM), Energy Dispersive X-ray Spectroscopy (EDX), and X-ray Diffraction (XRD). This project successfully synthesized CuO NPs through a green synthesis method by varying calcination temperatures. The UV-Vis spectra showed a maximum absorption peak at 368 nm for CuO-300 and CuO-500 NPs and 370 nm for CuO-400 NPs, corresponding to a band gap energy of 3.37 eV for CuO-300 and CuO-500 NPs, and 3.35 eV for CuO-400 NPs. The FT-IR spectra revealed prominent absorption bands at 552, 568, and 545 cm-1 for CuO-300, CuO-400, and CuO-500 NPs, respectively. In addition, the synthesized CuO-300, CuO-400, and CuO-500 NPs exhibited spherical morphology with differences in particle sizes ranging from 28.1 - 36.4 nm for CuO-300 NPs, 27.7 - 33.5 nm for CuO-400 NPs, and 27.3 - 34.5 nm for CuO-500 NPs. The EDX analysis showed only copper and oxygen elements present in the synthesized CuO NPs without any other impurity peaks. Lastly, all the synthesized CuO NPs exhibited a monoclinic structure, and the average crystallite size for CuO-300, CuO-400, and CuO-500 NPs was found to be 27.09, 25.22, and 25.14 nm, respectively
Evaluation of dynamic physicochemical property changes of purple sweet potato (Ipomoea batatas (L.) Lam) vinegar through stepwise fermentation
Get PDFVinegar is a health-beneficial condiment produced from two-stage fermentation involving alcohol conversion and subsequent acetification using high-sugar materials. However, vinegar production from starchy materials, such as purple sweet potato and cassava, has been rarely studied due to the limited conversion of starch to sugar. Since agarwood leaves are an underutilized by-product of agarwood cultivation, their infusion has been incorporated to enhance the physicochemical properties of vinegar, such as total phenolic content. This study aimed to evaluate the effects of different substrates and the presence of agarwood leaf infusion on the physicochemical and biochemical properties of vinegar production. Purple sweet potato and cassava obtained from the local market were saccharified using Aspergillus oryzae, followed by alcohol and vinegar fermentation. During this process, dynamic changes in the biochemical properties of vinegar were monitored, alongside phytochemical qualitative screening and organic acid analysis. Purple sweet potato showed a sugar content of 28.40 %wb, but the amino acid content after saccharification was lower compared to cassava. Vinegar incorporated with agarwood leaves displayed a similar trend in alcohol content as the water-treated samples. Dynamic changes in vinegars showed an increase in pH and a decrease in alcohol content during acetic acid fermentation. The total phenolic content of agarwood-treated vinegar was higher than that of water-treated vinegar. Quantitative analysis revealed that the acetic acid ranged from 1.176 mg/mL to 4.536 mg/mL for both vinegars. Besides, similar acetic acid production was observed in the water-treated and agarwood-treated vinegar at Day 10. The study showed higher potential for purple sweet potato for vinegar production. In addition, the incorporation of agarwood tea improved the biochemical properties of the vinegar while not affecting the alcohol content throughout the fermentation process and acetic acid level at the end of acetic acid fermentation
ExploreEasy: Smart and all-in-one trip management application
Get PDFAs travellers increasingly seek tailored and efficient experiences, current travel applications often fail to address diverse requirements and adapt to real-time changes. This research presents an AI-driven trip planning and recommendation system that employs a Hybrid Recommendation Algorithm, integrating Collaborative Filtering (CF) and Content-Based Filtering (CBF) to provide highly customised travel itineraries. Core components include automated accommodation suggestions, an inflation-aware budget estimation and management module that applies Jaccard Similarity and Weighted Averaging for accurate budget ranges, and a cost split feature for fair expense sharing among travellers. The system also provides real-time budget alerts to enhance financial transparency and control. To improve travel efficiency, the system incorporates intelligent route optimisation using the Travelling Salesman Problem (TSP), ensuring time-efficient and logically sequenced itineraries. Additionally, a similar-place substitution feature leveraging Geographic Filtering and Quality Thresholds increases flexibility by dynamically suggesting contextually relevant alternatives. Furthermore, the integration of real-time and extended weather forecasting enables dynamic itinerary modifications to enhance safety and adaptability. The system’s effectiveness was evaluated with real travel data, revealing significant improvements in personalisation, flexibility, financial confidence, and user satisfaction. By combining a hybrid recommendation engine with innovative features such as weather-aware itinerary adjustments, budget monitoring, expense splitting, and substitution-based adaptability, this project delivers a more intelligent, responsive, and user-centric trip planning experience than conventional platforms
Fundamental stock analysis with LLMs and qualitative data: Impact of government policies and economic trends
Get PDFThis report presents the development of a Virtual Analyst system for fundamental stock investment, powered by GPT-4o mini and other advanced technologies. The system leverages Large Language Models (LLMs) for processing and analysing qualitative data to provide comprehensive stock analysis and investment recommendations. The system integrates web scraping techniques to extract valuable information from diverse sources such as government policies, economic reports, news articles, and financial statements. The research process involved designing a modular architecture with five core components: financial report extraction, real-time news collection, inter-company relationship mapping, qualitative analysis of government policies and economic trends, and investment insight generation. Emphasis was placed on the qualitative analysis module, which leverages Retrieval-Augmented Generation (RAG) techniques to deliver contextually relevant insights. Preliminary testing validated the system's ability to generate accurate investment recommendations in JSON format. The conclusion highlights the system’s potential to democratize sophisticated financial tools and to empower retail investors with actionable insights into stock growth prospects. Planning for future work includes real-time data integration and scalability enhancements, ensuring alignment with the project’s objectives of transforming financial decision-making. The proposed methods and technologies have been justified as suitable for achieving the system’s objectives of delivering actionable and contextually relevant insights into stock growth prospects, thus demonstrating the potential to transform decision-making in fundamental stock analysis
Intellihire: An AI-powered interviewer for automated candidate selection
Get PDFThe recruitment process has gone a long way to determine the success of organisations in today's highly competitive job market. Traditional interview techniques conducted by human recruiters can be time-consuming, require a huge number of resources, and often suffer from scheduling challenges and inconsistent evaluation criteria, which can influence the way decisions are made. These struggles can cause an inconsistency in the way candidates are evaluated and ultimately result poor hiring decision making. Artificial intelligence (AI) is the up-and-coming technological process that addresses these problems in recruitment. IntelliHire: an AI Interviewer for automated candidate selection is a project that envisions building an extensive Audio-visual enabled machine understanding engine to automate the shortlisting from resumes till scoring interview sessions. IntelliHire provides an inexpensive, time-effective and unbiased way to replace traditional interview methods. This innovation attempts to minimize the time and resources expected from a recruitment process while improving precision in selection as well as providing fairness for job applicants. Ultimately, IntelliHire has the potential to revolutionize the hiring process, providing organizations with a powerful tool to make more informed and objective hiring decisions