Highlights in Web Engineering

Chapter 1

Understanding Beyond the Block: Overcoming Scalability Challenges for Blockchainpowered Healthcare IoT

By Ahmad Anwar Zainuddin [0000-0001-6822-0075], Amysha Qistina Amerolazuam, Nabilah Ahmad Nordin, Alin Farhain Abdul Rajat @ Abdul Razak, and Nur Alia Alina Abdul Rahman

Abstract – As blockchain technology is developing and finding its space in different economic domains, there is still an insufficient understanding of the limitations, abilities, and scope of the technology in the future. This chapter aims to address these rising needs of scalability challenges for blockchain-enabled healthcare IoT by implementing blockchain technology to securely and efficiently share healthcare data within the healthcare and brain industries systems to ensure its efficient and widespread implementation. Furthermore, the traditional practice of health care in the health care and brain domain is changed by blockchain for further reliability and assurance of proper diagnosis and treatment securely by sharing patient details. Consequently, it is established that blockchain technology is highly expected to improve the personal, authentic, and secure healthcare system to keep track of extensive real-time clinical information of the patient’s health and place that information or data in a secure frame and keep updating that on the blockchain.

Cite this chapter as:

Zainuddin, A. A., Amerolazuam, A. Q., Nordin, N. A., Abdul Rajat @ Abdul Razak, A. F., & Abdul Rahman, N. A. A. (2025). Understanding Beyond the Block: Overcoming Scalability Challenges for Blockchain-powered Healthcare IoT. In S.-
C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE)(pp. 3-30). MMU Press.

Chapter 2

Empowering Health: Revolutionising Diabetes Prediction with Cutting-edge Machine Learning Web Applications

By Rakesh Bhavsar, Vishvjit Thakar and Hardik M Patel

Abstract – Diabetes is an ongoing medical disorder marked by high blood glucose levels. Severe complications may involve heart disease, high blood pressure and visual issues. Timely medical intervention and lifestyle modifications are contingent upon early identification. With the promise to enhance outcomes and optimise care, machine learning approaches have emerged as formidable tools for tailoring treatment strategies based on real patient data and predicting diabetes. The various methodologies to be applied in diverse models entail K-Nearest Neighbour (KNN), Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), and Random Forest (RF). With these models, it is possible to predict if a person has diabetes by identifying the relevant signs. In order to deliver current forecasts based on factors that cause diabetes, we will incorporate the models that were taught into a web application., such as body mass index, age, and insulin levels. We utilised the Kaggle dataset to construct a machine learning-based model for predicting diabetes. We developed a Flask application for diabetes prediction to provide insights into health status and risk factors. We organised the app into modules, such as routes, templates, forms, and static assets, and used Flask’s modular structure. Our site uses HyperText Markup Language (HTML), Cascading Style Sheets (CSS) and JavaScript to create dynamic content and enable interactivity. Machine learning generates a diabetes risk score based on personal details. We have deployed this algorithm to the web with Flask (a Python framework). Predictive models in healthcare could lead to better patient outcomes. Health providers might employ them through predictive analytics within electronic medical records, mobile apps for monitoring individuals’ wellness statuses, or population health management systems that aim to identify those most in need so that interventions can be prioritised while keeping track over time. It would be essential for future research efforts to integrate various data sources if we were to obtain more accurate results.

Cite this chapter as:

Bhavsar, R., Thakar, V., & M Patel, H. (2025). Empowering Health: Revolutionising Diabetes Prediction with Cutting-edge Machine Learning Web Applications. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE) (pp. 31-58). MMU Press.

Chapter 3

Wireless Portability in Focus: Challenges and Opportunities for the Next Decade

By S. Kannadhasan, S. Syed Jamaesha, C. Madumalar and T. Santhiya

Abstract – The increasing adoption of wireless communication technology has broadened the scope and utility of devices, allowing access from almost any location. Educational institutions that traditionally depended on wired devices for administrative tasks, might benefit significantly from wireless technology. The study employs a case study approach to investigate how wireless communication technologies might improve administrative tasks in learning environments. While many businesses still rely on wired communication, switching to wireless technology may resolve some inefficiencies. The findings suggest that wireless technology can reduce operational costs and improve flexibility in administrative operations, though concerns regarding data security remain. The paper highlights the long-term advantages of wireless technology while addressing concerns like data transfer security. It also discusses the unpredictable nature of IT breakthroughs and forecasts the industry’s ongoing growth in wireless communication.

Cite this chapter as:

Kannadhasan, S., Syed Jamaesha, S., Madumalar, C. & Santhiya, T. (2025). Wireless Portability in Focus: Challenges and Opportunities for the Next Decade. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering
(HIWE) (pp. 59-77). MMU Press.

Chapter 4

Multi-Optimisation Approaches for Efficient Wireless Communication Systems

By S. Kannadhasan, P. Palaniayammal, G. Priya and P. Sneka

Abstract – Co-site interference with electronic car systems is on the rise. Communication system designers face challenges due to the flexibility of using multiple channels in crowded electromagnetic settings and the competing demands for high performance. Thus, we must devise methods for lowering electromagnetic interference in complex systems to guarantee dependable satellite and VHF communications. In this work, we provide a method to identify interference and suggest a mitigation approach in mobile Satellite Communication (SatCom) networks with co-site VHF systems. We simulate validating the interference mode and enhancing the suppression process. The latest experimental findings demonstrate the effectiveness of the suppression strategy in limiting the propagation directions of interferers. The ground station of a SatCom system often causes electromagnetic interference (EMI) when exposed to the external electromagnetic field. This study highlights an uncommon yet serious issue with EMI. SatCom produces EMI, which other wireless systems need to protect themselves from. However, grounding or insulating the emission source or affected component is the most used technique for reducing EMI from radio systems. This study discusses the method for lowering EMI in transmission channels. This might improve the performance of Electro-Magnetic Compatibility (EMC) regulation for highly mobile, always-connected SatCom systems.

Cite this chapter as:

Kannadhasan, S., Palaniayammal, P., Priya, G. & Sneka, P. (2025). MultiOptimisation Approaches for Efficient Wireless Communication Systems. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE)
(pp. 79-95). MMU Press.

Chapter 5

BERT-Based Machine Learning Model for Classifying Emotions

By Sellappan Palaniappan, Rajasvaran Logeswaran and Saw Yi Jiau

Abstract – We communicate and interact with people both face to face and remotely using words. In today’s digitally connected world, we communicate with people from all over the world through social media such as Facebook, email, twitter, blogs, and mobile devices. Understanding peoples’ emotional states (happy, angry, sad, and so on) via their spoken or written words can help us relate to them more appropriately. In this study, we use a BERT-based machine learning model to classify people’s emotions using their written words. The model has potential applications in education, healthcare, marketing, and workplace settings. Our study focuses on six types or categories of emotions: joy, gratitude, love, anger, sadness, and fear. Our study used a combination of both synthetic and real-world data to train and test our model. We implemented the model using Python’s PyTorch and Hugging Face Transformers library and achieved an overall accuracy of 80% with fair performance in all six emotion categories. We used metrics like precision, recall, and F1-score to assess its effectiveness. By combining both synthetic and real-world data, our study overcomes the problem of data availability, balance and representativeness. While our model has potential applications, it has several limitations such as presence of biases arising from the use of synthetic data, not including other emotion types, not including spoken words, not including other languages, not addressing the intentions of people behind their written words. We also did not address privacy and ethical issues related to classifying people’s emotions or revealing their emotional states to others.

Cite this chapter as:

Palaniappan, S., Logeswaran, R., & Jiau, S. Y. (2025). BERT-Based Machine Learning Model for Classifying Emotions. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE) (pp. 97-114). MMU Press.

Chapter 6

Next-Generation Virtual Environment Training for Cybersecurity in the Maritime Logistics Industry: A Position Paper

By Cheng Yun-Quan, Rory Hopcraft, Kimberly Tam and Chin Ji-Jian

Abstract – The maritime logistics industry is the most dependent on it in the modern world, yet it is not well protected against cyber risks. Recently, the maritime industry has been transitioning into industry-specific tech, and threats of cyberattacks are increasing as the adoption of technology and automation continues. One of the most effective methods to protect industries and ships from such security incidents, is to train seafarers to mitigate and handle such incidents. However, to the best of our knowledge, we did not find anything that addresses this gap. This article will present the idea and benefits of using virtual environments for next-generation maritime skill development and its ideal method of executing it. We will be discussing the use of virtual reality, the factors behind security incidents, its syllabus design philosophy, target users, and experiments. The purpose of this article is to serve as a preliminary research and outline while positioning our project titled: Next generation of maritime skill development: future problems and technologies. This positioning paper could produce a work that can be the test bed and platform to guide other projects and for future similar projects to be built upon, with the hope of seeing future works dive deeper into the specifics of virtual training program research, regardless of discipline.

Cite this chapter as:

Yun-Quan, C. Y.-Q., Hopcraft, R., Tam, K., & Ji-Jian, C. (2025). Next-Generation Virtual Environment Training for Cybersecurity in the Maritime Logistics Industry: A Position Paper. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE) (pp. 115-136). MMU Press.

Chapter 7

Software Engineering Challenges and Solutions in AI-Driven Systems

By S. Kannadhasan, S. Ganesh and SMK Ramakrishnan

Abstract – Software engineering has changed paradigm due to the rapid development of artificial intelligence (AI) technologies, posing new problems that call for creative solutions. This study examines the many obstacles that arise when creating AI-driven systems, such as bias and data quality, the interpretability of models, integration with current systems, and ethical issues. It discusses the impact of these difficulties on software design, testing procedures, and deployment tactics. Moreover, we provide a variety of remedies, including establishing strong data governance frameworks, putting understandable AI strategies into reality, using continuous integration and deployment procedures, and encouraging multidisciplinary cooperation. By addressing these issues and implementing practical solutions, software developers can enhance the dependability, equity, and overall effectiveness of AI-driven systems, thereby paving the way for more ethical and sustainable AI development. Training using Large Language Models (LLMs) has shown impressive results in several domains. One such use is providing accurate symptom descriptions to support medical diagnosis. This study examines the domain of type sickness diagnosis, a traditional usage of AI in medicine, to investigate the valuable use of LLMs in healthcare. Deep network-driven image classification models, such as ResNet, VGG, DenseNet, etc., are the foundation of traditional AI-based methods for detecting various illnesses. These models often only provide one-dimensional capabilities and lack mechanical understanding. The program analyses user-submitted photos using sophisticated machine learning techniques and offers prompt input on dermatological concerns. The primary objective is to improve early intervention and accessibility to dermatological treatment, especially for underprivileged populations.

Cite this chapter as:

Kannadhasan, S., Ganesh, S., & Ramakrishnan, SMK. (2025). Software Engineering Challenges and Solutions in AI-driven Systems. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE) (pp. 137-152). MMU Press

Chapter 8

Credit Card Fraud Detection Using Deep Learning: Enhancing Accuracy and Efficiency

By S. Ganesh, S. Kannadhasan and M. Selvapradap

Abstract – The usage of credit cards, especially MasterCard, has significantly increased in recent years due to the growth of online purchases. However, this surge has coincided with an increase in fraudulent activities. Fraudulent transactions, such as those using stolen or fraudulent credit cards, now impact financial institutions worldwide. It is now challenging to identify and stop these fraudulent transactions. Nowadays, criminals use more complex strategies to trick victims and financial systems, which makes it harder to spot fraud of this kind. By developing an automated method to identify fraudulent MasterCard transactions using cutting-edge machine learning techniques, the suggested solution seeks to solve this rising threat. In particular, the system analyses transactional data using decision tree and random forest techniques. These algorithms work well for seeing trends in big data sets and determining whether a transaction is fraudulent. The initial stage in developing the system is to gather transaction data, which includes elements like transaction amount, time, and client information. We will separate this data into training and testing sets to develop a model. Big datasets particularly benefit from an ensemble learning technique known as Random Forest. To increase prediction accuracy and lower the chance of overfitting, it builds many decision trees during training and outputs the majority vote for classification. By graphically depicting the decision-making process, the Decision Tree algorithm further improves the interpretability of the model. It makes it simpler for analysts to comprehend and act upon the findings. With an accuracy rate of 98.6%, the system has shown excellent performance, demonstrating its ability to detect fraudulent transactions. Furthermore, data visualisations and other visual depictions of the outcomes assist stakeholders in rapidly understanding the model’s efficacy. The solution saves time and money by automating the detection process, giving financial institutions a valuable tool to fight fraud effectively. Customers and financial institutions benefit from this strategy’s improved fraud detection and increased online transaction security.

Cite this chapter as:

Ganesh, S., Kannadhasan, S., & Selvapradap, M. (2025). Credit Card Fraud Detection Using Deep Learning: Enhancing Accuracy and Efficiency. In S.-C. Haw, S. Y. Ooi, Y. J. Chew, & J. J (Eds.), Highlights in Web Engineering (HIWE) (pp.153-175). MMU Press.