Journal of Smart Science and Technology

Effect of slag powder on the performance of self-compacting concrete

Shen Si — 2026-09-30

This study employs S95-grade slag powder as an additive and formulates four distinct groups of C45 self-compacting concrete with a fixed fly ash content of 20% and slag powder proportions of 0%, 10%, 20%, and 30%, respectively. The effects of these on workability and mechanical properties such as compressive strength, splitting tensile strength and flexural strength are systematically studied. The results show that slag powder can improve the filling capacity and gap permeability of concrete, but the segregation rate increases with the increase in dosage. The comprehensive performance is optimal when the slag powder dosage is 20%. In terms of mechanical properties, the optimal strength at all ages was achieved when 20% slag powder was added, with a 90-day compressive strength of 62.4 MPa and later strength better than the benchmark group. In summary, a 20% addition of slag powder can achieve a favorable balance of multiple properties in self-compacting concrete, providing technical support for its engineering applications.

Do pollutant interaction terms improve water quality index classification? An empirical study using support vector machine and random forest on Malaysian river data

Najihah Abdol Raof — 2026-09-30

This study examines whether engineered interaction effects among the six main parameters of the Malaysian Water Quality Index (WQI) can improve multi-class water quality classification. River monitoring data from the Department of Environment from 2020 to 2022 were pre-processed and statistically validated prior to model development. Six interaction terms were constructed and evaluated using Support Vector Machine (SVM) and Random Forest (RF) classifiers. Model training was performed using repeated 10-fold cross-validation with Cohen’s Kappa as the optimisation criterion, followed by evaluation on the testing set data. Results showed that models built using only the six main WQI parameters consistently outperformed those incorporating interaction terms. The SVM main effects model achieved the highest performance, with 95.02 per cent accuracy and a Kappa value of 0.9105, indicating an excellent multi-class reliability. Interaction-based models produced a lower Kappa value, suggesting that both SVM and RF were already able to capture complex non-linear relationships without the need for engineered interaction terms. Class-level metrics further demonstrated strong performance across major WQI classes, while reduced sensitivity in Class IV reflected severe class imbalance. Overall, the findings indicate that explicit interaction effects do not improve WQI prediction and that a parsimonious model using only the six WQI parameters is sufficient for reliable classification.

Genetic diversity of sharks in Malaysia and Southeast Asia (SEA) region: A secondary data approach

Izzati Adilah Azmir — 2026-09-30

Sharks are essential to the balance and well-being of marine ecosystems. However, shark populations are declining globally, raising significant conservation concerns, particularly the loss of genetic diversity, which threatens the species' ability to adapt over time. This study aimed to collect CO1 gene sequences of shark species from the public database GenBank and the Barcode of Life Data (BOLD) system. The objective of the study was to establish a phylogenetic tree, and to investigate the genetic diversity of sharks in Malaysia and the Southeast Asia (SEA) region. A total of 83 sequences from 43 species, 24 genera, 16 families, and six orders were found in Malaysia. The phylogenetic analysis comprised the Malaysian shark sequences and reference sequences from the SEA region using Neighbour-joining (NJ) and Maximum Likelihood (ML) methods, with the NJ tree constructed under the Kimura 2-Parameter (K2P) model and the ML tree inferred using the GTR + G + I model. Nodal support was evaluated using 1,000 bootstrap replicates. Notably, species such as Hemigaleus microstoma and several Carcharhinus species exhibited low genetic diversity (Hd = 0.5000; π = 0.0011), while Squalus altipinnis displayed the highest genetic diversity (Hd = 1.000 and π = 0.0603). These findings offer critical baseline genetic data for Malaysian and SEA shark populations, emphasizing species that may be particularly vulnerable to anthropogenic pressures, which can contribute to conservation planning, species management, and regional monitoring strategies aimed at ensuring the long-term viability of shark populations in Malaysian and Southeast Asian waters.

Comparative evaluation of supervised machine learning algorithms for DDoS detection

Muhammad Ammar Ahmad Nazam — 2026-09-30

Maintaining the reliability, availability, and security of modern networked systems is significantly challenged by the distributed denial-of-service (DDoS) attack. Traditional defence mechanisms often struggle to detect evolving attack patterns. This issue highlights the need for an advanced and structured machine learning approach. To address these issues, this study analyses DDoS attacks and evaluates performance using supervised machine learning techniques. A structured methodology consisting of four phases: - requirements gathering, design, development, and testing was adopted. During the requirements phase, related studies were reviewed, and the CICDDoS2019 dataset was selected. In the design phase, essential preprocessing tasks were performed, including handling missing values, removing duplicate data, performing feature selection, and balancing the dataset. Several supervised machine learning strategies were applied in this study. These comprised decision-tree ensembles, linear classifiers, neighbour-based prediction models, gradient boosting approaches, probabilistic classifiers, and support vector methods, which were applied in the code development phase. The selected algorithms were trained using the dataset. During the testing phase, multiple indicators were employed to examine how well the models performed. The assessment considered correct prediction rates, the accuracy of positive classifications, coverage of true positive instances, and the combined measure of precision and recall represented by the F1-score. The results show that tree-based methods outperformed the other models. Hierarchical decision-making using trees, ensemble predictions through random forests, and gradient boosting techniques all achieved highly accurate detection performance. Furthermore, confusion matrix analysis was conducted to examine the detailed performance of each model in classifying instances. Overall, the findings indicate that supervised learning algorithms perform well and can facilitate the scalable implementation of cybersecurity systems.

Water quality dynamics and operational stability of a campus-scale aquaponic system

Nursalwani Sabdani — 2026-09-30

Aquaponics is increasingly recognized as a sustainable food production system due to its efficient use of water and nutrients through the integration of aquaculture and hydroponics. This case study evaluates the water quality dynamics and operational stability of a Deep-Water Culture (DWC) aquaponic system at the Eco Melati Garden at one of the higher education institutions in Malaysia. Pak Choy (Brassica rapa var. chinensis) and red tilapia were cultivated over a five-week production cycle. Key water quality parameters, including pH, ammonia, nitrite, and nitrate, were monitored weekly using standard test kits and assessed based on acceptable ranges reported in the literature. Operational stability was evaluated using a simplified scoring framework based on daily observations of fish health, water circulation, and water clarity. Results showed that pH remained within 7.0 to 7.6, and ammonia ranged from 0.25 mg L^-1 to 0.50 mg L^-1, nitrite level remained near zero, and nitrate varied between 5 mg L^-1 to 20 mg L^-1. These results were proven within acceptable ranges reported in the literature. Fish health and circulation achieved high stability scores while turbidity fluctuations indicated opportunities for better maintenance. Overall, this study demonstrated stable operational performance and maintained suitable environmental conditions for warm-water fish as well as leafy vegetables. This short‑term study highlights the feasibility of campus‑scale aquaponics while identifying limitations and operational improvements for future long-term monitoring.

Adoption of solar energy to reduce carbon footprint for the Trinidad and Tobago residential sector

Marlon Farmer — 2026-09-30

Climate change poses significant risks globally, particularly for Small Island Developing States (SIDS) such as Trinidad and Tobago (TT), which are highly dependent on fossil fuels for power generation. The country emits substantial greenhouse gases, with power generation accounting for 27% of total carbon dioxide (CO₂) emissions. Residential electricity consumption accounts for approximately 31% of this demand. This figure has grown due to post-COVID-19 shifts in household energy use, such as remote work, online education, increased home-based business, and medical operations. These changes have heightened the need for reliable and sustainable energy sources. This study aims to design, optimise, and evaluate a renewable energy (RE) system suitable for domestic households in TT using photovoltaic (PV) technology. An annual energy consumption profile of 15,739 kWh with associated CO₂ emissions of 9,947 kg was used as the base case in the HomerPro software. Sensitivity analyses were conducted on various subsidised and unsubsidised electricity rates, including 0.05, 0.12, 0.22, and 0.35 USD per kWh, assessing Net Present Cost (NPC), Levelized Cost of Energy (LCOE), operating cost, and annual CO₂ emissions. The results indicate that at current subsidised and modest unsubsidised rates, RE systems are not economically favourable. However, at 0.15 USD per kWh, a 10% RE penetration becomes viable. When modelled at the regional average unsubsidised rate of 0.22 USD per kWh, renewable integration becomes more cost‑effective, achieving a 36% penetration and reducing household CO₂ emissions by 422 kg annually. These findings highlight the potential for economically and environmentally sustainable residential energy transitions through targeted policy and financial frameworks.

Integrating aquaponic systems into higher education: A comprehensive evaluation of the UiTM Eco Melati Garden living lab for sustainability learning

Noor Hafiza Mat Noor — 2026-09-30

Living labs based on aquaponics are increasingly recognised as successful sustainability education tools in universities that help to improve environmental awareness, practical knowledge, and participation of the community. This study explored the Eco Melati Garden aquaponic system at UiTM Shah Alam, a student‑driven project that started as a hydroponic project and then grew into a living lab for sustainable food production, closed-loop resource management, and experiential learning. The purpose of the study is to measure students' sustainability skills, experiences, involvement, community impact, and environmental, social, and governance (ESG) awareness. A structured questionnaire was administered to 191 students, and the data were analysed across five domains. Findings revealed that sustainability literacy (mean = 3.42), experiential learning (mean = 3.33), active engagement (mean = 3.15), awareness of food security and community impact (mean = 3.14), and positive ESG awareness (mean = 3.23) are favourable. In general, the Eco Melati Garden system is a successful living lab, as it improves students' skills, facilitates SDG-oriented learning, and fosters sustainable culture which can be a model for sustainable education in Malaysian institutes of higher education.

Shaping sustainable agriculture with hydrogels: A bibliometric analysis of advances and research frontiers

Alzuveny Wise Jinual — 2026-09-30

This study outlines the research fields regarding hydrogel applications in agriculture, which include water-retention materials and multifunctional systems for soil enhancement, nutrient management and ecological remediation. Keyword co-occurrence analysis was performed using VOSviewer with a minimum threshold of five occurrences and a cluster size of seven using Scopus database data (2019–2025), resulting in six major thematic clusters. The findings show a significant increase in related publications since 2019, with China being the primary contributor. Six interconnected clusters highlight research frontiers in agro‑physiology and water-use management, biocompatibility and bio-interfaces, adsorptive remediation, water retention and sustainable agriculture, hydrogel engineering, and soil remediation. Eco-functional hydrogels that support climate-resilient and resource-efficient farming are emerging from these clusters of materials science and agricultural sustainability. The study provides valuable insights into global collaborative networks and thematic evolution, despite its limitation to Scopus-indexed English publications. This study's visualisation and understanding of hydrogel-related knowledge structures and biodegradable hydrogels affects sustainable agriculture and circular bioeconomy in a novel way.

A review of digital learning tools in smart campus environments

Azyan Yusra Kapi — 2026-09-30

Smart campus development has attracted significant attention in recent years as higher education institutions adopt digital technologies to support teaching, learning, and campus operations. The integration of digital learning tools within smart campus environments aims to improve learning accessibility, management efficiency, and the overall student experience. However, existing studies on digital learning tools are often reported in isolation, limiting a consolidated understanding of how these technologies function collectively within smart campus ecosystems. This paper adopts a narrative review approach supported by thematic synthesis to examine several categories of digital learning tools commonly discussed in smart campus contexts, including mobile learning platforms, intelligent tutoring systems, learning analytics, augmented and virtual reality applications, and campus support systems. The review synthesises recurring patterns from the literature by examining the functional roles, educational applications, implementation challenges, and smart campus contributions of these tools. The findings highlight the growing interconnection between digital learning technologies and broader smart campus infrastructures, particularly in relation to accessibility, personalised learning, data-driven decision-making, sustainability-related outcomes, energy optimisation, space utilisation, and resource efficiency. The paper concludes by identifying future directions related to integrated digital systems, methodological clarity in smart campus research, and human-centred approaches in smart campus design.

ANN-based predictive classification of eco-friendly microwave absorbers for C-band frequency applications

Azizah Ahmad — 2026-09-30

Eco-friendly microwave absorbers derived from agricultural waste have gained increasing attention due to their environmental benefits and potential for high electromagnetic absorption performance. However, predicting their multiband characteristics remains challenging because of the nonlinear interaction between material properties and electromagnetic waves. In the present study, an Artificial Neural Network (ANN)-based predictive classification framework is proposed to assess the absorption performance of eco-friendly microwave absorbers operating in the C-band frequency range (4-8 GHz). A dataset consisting of 342 experimentally measured absorption samples categorised into three slot size groups (small, medium and big) was pre-processed using boxplot-based outlier removal and min-max normalisation to improve data quality. An ANN Regression model was then developed to analyse the correlation between input (frequency) and output (absorption) across the C-band. The regression analysis generated high R-values for the training, validation and testing datasets, demonstrating strong predictive consistency across all data partitions. The Multilayer Perceptron (MLP) architecture was then trained using three algorithms, namely Levenberg-Marquardt (LM), Resilient Backpropagation (RBP) and Scaled Conjugate Gradient (SCG) to classify absorption performance, with datasets partitioned into 70% training, 15% validation and 15% testing. Among the selected configurations, RBP achieved the lowest mean squared error while maintaining the same classification performance as LM and SCG within the evaluated dataset. These findings demonstrate the potential of ANN modelling as a practical approach for predicting and classifying the C-band absorption performance of eco-friendly microwave absorbers, supporting efficient absorber design for future RF applications.