Forthcoming Articles
International Journal of Information and Communication Technology
Forthcoming articles have been peer-reviewed and accepted for publication but are pending final changes, are not yet published and may not appear here in their final order of publication until they are assigned to issues. Therefore, the content conforms to our standards but the presentation (e.g. typesetting and proof-reading) is not necessarily up to the Inderscience standard. Additionally, titles, authors, abstracts and keywords may change before publication. Articles will not be published until the final proofs are validated by their authors.
Forthcoming articles must be purchased for the purposes of research, teaching and private study only. These articles can be cited using the expression "in press". For example: Smith, J. (in press). Article Title. Journal Title.
Articles marked with this shopping trolley icon are available for purchase - click on the icon to send an email request to purchase.
Online First articles are also listed here. Online First articles are fully citeable, complete with a DOI. They can be cited, read, and downloaded. Online First articles are published as Open Access (OA) articles to make the latest research available as early as possible.
Articles marked with this Open Access icon are Online First articles. They are freely available and openly accessible to all without any restriction except the ones stated in their respective CC licenses.Register for our alerting service, which notifies you by email when new issues are published online.
International Journal of Information and Communication Technology (37 papers in press) Regular Issues
Abstract: The icing of pillar insulators in actual substation operating environments can lead to potential hazards such as power outages. Effective deicing of pillar insulators is of great significance. As an emerging deicing method, jet heating deicing lacks existing models. This study innovatively explores the temperature field distribution through advanced numerical simulation methods based on COMSOL, unlike traditional methods that rely mainly on experimental measurements or simple theoretical models. It further describes the process of establishing a three-dimensional model in detail. Through simulation analysis, the three-dimensional temperature field distributions of post insulators under different working conditions or heat source parameters are obtained, visually presenting the change trends and distribution characteristics of temperature. The research results provide a theoretical basis for in-depth understanding of the heat transfer mechanism during the deicing process of post insulators. Keywords: pillar insulator; three-dimensional temperature field distribution; distribution characteristics; heat transfer mechanism. DOI: 10.1504/IJICT.2025.10074671
Abstract: Multilingual classroom contexts pose significant pedagogical challenges, particularly when students have diverse native languages and varying levels of Korean proficiency. This study introduces a data-driven instructional model for Korean translation education that employs machine learning to address learner diversity. The model evaluates translation outputs, identifies learner-specific error patterns, and personalises instruction based on three key variables: native language influence, historical translation accuracy, and individual learning progression. A dataset of Korean translation tasks was collected from university students representing six L1 backgrounds Chinese, Vietnamese, Arabic, Russian, Japanese, and Spanish. Texts were pre-processed through tokenisation, lemmatisation, and POS tagging, with Word2Vec embeddings used for feature extraction. The proposed Sparrow Search Optimiser Tuned Attention-based Sequence-to-Sequence (SSO-Attn-Seq2Seq) model demonstrated substantial improvements, achieving 8891% across accuracy, precision, recall, and F1-score. Results highlight its adaptability in handling idiomatic expressions and syntactic variation, providing a scalable solution for multilingual Korean language education. Keywords: multilingual classroom settings; grammatical variations; languages; SSO-Attn-Seq2Seq. DOI: 10.1504/IJICT.2025.10074672
Abstract: Music serves as a convenient and effective tool for emotional regulation and holds significant value in psychological adaptation. Addressing the issue that existing research has overlooked real-time changes in college students interests, leading to insufficient analysis of the impact on psychological adaptation, this study first embeds music input into a long short-term memory network model, modelling the sequence processing issue as the Markov decision process, and uses a multilayer perceptron model as the decision agent. High and low-score decision actions are input into a pseudo-twin network to generate delayed rewards. The agent gradually learns strategies that maximise rewards, enabling reliable music recommendations. Finally, the study analyses the systems impact on college students psychological adaptation. Experimental results show that the proposed models hit rate improves by at least 4.73%, significantly enhancing college students psychological well-being. Keywords: psychological adjustment; reinforcement learning; music recommendation; Markov decision process; long short-term memory network model. DOI: 10.1504/IJICT.2025.10074808
Abstract: This study aims to enhance traditional management accounting by developing an AI-based system for cost accounting and budget optimisation. The proposed framework follows a structured nine-step process, beginning with problem identification and concluding with system validation. Each stage ensures transparency and effective implementation. AI contributes to improved prediction accuracy, cost reduction, and more reliable financial decision-making, while highlighting the limitations of outdated, paper-based methods. In practice, AI assists in tasks such as tax processing, error detection, and forecasting. Historical data are used to train AI models, which are then applied to accounting operations and validated for accuracy and relevance. Despite challenges in integration, scalability, and ethical considerations, results indicate strong reliability, with Cronbachs alpha and composite reliability values exceeding 0.8 in SEM tests. Overall, the AI model outperformed traditional methods by reducing costs and adapting effectively to workload variations. Keywords: AI-driven; management information system; MIS; cost accounting; budget optimisation; machine learning; decision support systems; financial management; predictive analytics; reinforcement learning; digital transformation. DOI: 10.1504/IJICT.2025.10074809
Abstract: To address the challenges of information overload and resource misallocation in the tourism industry, this paper proposes an intelligent service framework that integrates multi-source big data with knowledge graphs. By constructing a tourism-specific knowledge graph from the Yelp dataset (containing over 12,537 POIs and 45,821 users) and combining relational graph convolutional networks with long short-term memory models, the framework achieves precise personalised recommendations and dynamic resource optimisation. The proposed multi-task learning architecture jointly optimises recommendation accuracy and resource prediction performance. Extensive experiments show that the model significantly outperforms baseline methods, achieving a Precision@10 of 0.0914 and Recall@20 of 0.2542, along with a 21.73 root mean square error in flow prediction demonstrating notable improvements in interpretability and robustness. This study provides an effective technical pathway for enhancing tourism service intelligence and operational efficiency. Keywords: knowledge graph; smart tourism; resource optimisation; recommendation system; big data analysis. DOI: 10.1504/IJICT.2025.10074810
Abstract: This paper proposes a prediction model based on a spatio-temporal graph convolutional network to capture the spatio-temporal dependencies of user interactions and improve the accuracy of predicting trends in the dissemination of ideological and political themes. Experimental results show that this method achieves an average improvement of 6.8% in the F1 score and a reduction of 7.2% in the prediction root mean square error for key metrics such as the probability of ideological and political hot topics appearing in the coming week and changes in regional sentiment distribution. This model effectively integrates the spatial structural information of social networks with dynamic temporal features, providing a reliable computational tool for quantitative analysis and forward-looking assessment of ideological and political dynamics in social media environments. These aid relevant departments in timely sensing and guiding the online ideological and political ecosystem. Keywords: spatio-temporal convolutional networks; social media platforms; ideological and political communication. DOI: 10.1504/IJICT.2025.10074811
Abstract: This paper suggests a sentiment monitoring system in the preschool industry to monitor the sentiment of the people regarding early childhood learning. The system gathers and pre-processes social media posts with help of transformer-based language models and entropy scoring, sentiment classification, and unpredictability measurement. The information is collected and presented in real-time on a dynamic dashboard. Findings indicate that there is no consistency between the magnitude and the sentiment change of post volume and that entropy-based metrics provide a more precise analysis of the volume. The system is capable of identifying any abrupt shifts in the mood and thus organisations can be able to respond to current issues at the earliest opportunity. In preschool learning, this method increases parent involvement, organisational sensitivity, and relationship development by using AI-based sentiment analysis. Keywords: sentiment analysis; social media monitoring; emotional entropy; transformer models. DOI: 10.1504/IJICT.2025.10074812
Abstract: Currently, artistic images are scarce with limited sample sizes, and most sentiment analysis relies on low-level image features with low accuracy. To address this, this paper first extracts two-dimensional features from images in different colour spaces. It then employs multi-scale convolutional kernels to extract deep semantic information from images, fusing feature information from different dimensions to effectively preserve semantic features across scales. Finally, the transfer component analysis algorithm is employed to reduce dimensionality of features in source and target domains within original space. An improved joint subspace learning method is used to learn a feature transformation subspace, reducing the conditional probability distribution distance between source and target domains while balancing recognition accuracy across categories. Model optimisation is achieved through adversarial training. Experimental results demonstrate that the proposed model improves recognition accuracy by at least 3.82%, effectively enhancing the accuracy of emotional recognition in artistic images. Keywords: artistic image emotion recognition; feature fusion; transfer learning; adversarial training; feature extraction. DOI: 10.1504/IJICT.2025.10074813
Abstract: Automatic composition struggles to balance multi-theme planning with strict polyphonic constraints. To address this, this paper proposes an adaptive genetic framework for multi-topic polyphonic music generation. In the scheme, first a bar-aligned, voice-aware representation is prepared with tonal cues and a theme schedule. Then a domain-aware evolutionary core explores the search space through voice-preserving crossover, musically constrained mutation, and lightweight local repair around exposed phrases. Finally, a composite evaluator guides selection while an adaptive controller adjusts operator rates using diversity and stagnation signals. Experiments on chorale, chamber, and modern tonal sets show fewer rule violations, higher consonance with 83% vertical consonance and tonal stability, stronger theme recognisability, and faster convergence without extra runtime. The approach delivers structured, stylistically credible music with strong controllability, clear diagnostics, and room for interactive use. Keywords: algorithmic composition; polyphonic music; adaptive genetic algorithm; thematic scheduling. DOI: 10.1504/IJICT.2025.10074814
Abstract: The rapid expansion of the digital economy heightens the need for privacy and trust in intellectual property transactions. Traditional centralised approaches to identifying legal conflicts in intellectual property contracts are prone to data leakage and fail to balance transparency with confidentiality. This paper proposes a self-identification method for legal conflicts in intellectual property contracts using zero-knowledge proofs. By combining a light gradient boosting machine learning model with the zero-knowledge succinct non-interactive argument of knowledge protocol, our approach allows verifiable detection of potential legal conflicts without revealing sensitive information. Experiments on the United States patent and trademark office patent dataset demonstrate that the method achieves high performance in conflict prediction (area under the receiver operating characteristic curve = 0.872) and verification efficiency (<10 ms), providing a novel and practical framework for privacy-aware legal technology. Keywords: zero-knowledge proof; ZKP; intellectual property contract; automatic identification of legal conflicts; privacy protection; machine learning. DOI: 10.1504/IJICT.2025.10074815
Abstract: To address the critical global demand from 1.3 billion English as a Foreign Language learner for personalised reading materials, this study develops a dual-channel regularised variational autoencoder. The model systematically overcomes conventional limitations in readability control and semantic coherence by establishing dynamic mappings between educational linguistic features and latent space, designing a novel readability-driven regularisation loss that integrates lexical complexity, syntactic simplification, and discourse cohesion, and implementing curriculum learning for progressive optimisation. Comprehensive evaluations on the Newsela benchmark corpus demonstrate statistically significant improvements: 7.2% in BLEU-4, 32.8% reduction in readability errors, and 20.6% enhancement in teacher-assessed quality. This framework provides an efficient solution for adaptive learning systems, advancing intelligent generation and scalable deployment of educational resources with high practical utility. Keywords: optimised variational autoencoder; English reading text generation; readability control; integration of educational features; Newsela dataset. DOI: 10.1504/IJICT.2025.10074816
Abstract: Green power consumption has become a key challenge in the energy transition. Existing research struggles to capture complex relationships between the spatio-temporal dynamics of the power system and policy interventions. To this end, this paper first designs a power load forecasting model based on spatio-temporal graph convolutional networks. The model dynamically adjusts the graph structure according to users electricity consumption patterns and introduces a weighted skip connection mechanism, assigning different weights to connections at different time steps. Then, a mathematical model for optimal combinations of power consumption policies is established. Through deep reinforcement learning algorithms interacting with the environment, it solves for the optimal combination of power consumption policies that minimise economic and carbon emission costs. Experimental outcome demonstrates that the proposed method achieves a green power consumption rate of 97.16%, outperforming comparison methods, thus helping to promote efficient green power consumption. Keywords: green power; consumption policy; spatio-temporal graph convolutional network; deep reinforcement learning algorithms; skip connections. DOI: 10.1504/IJICT.2025.10074817
Abstract: Rapid urbanisation intensifies pressures on urban landscapes, driving sustainability challenges like ecological degradation and unequal green space access. This study develops a genetic algorithm (GA)-based multi-objective optimisation (MOO) framework for sustainable landscape planning in Nanchang Citys first ring road. The non-dominated sorting genetic algorithm II (NSGA-II) is adopted to simultaneously optimise ecological, social and economic objectives. Spatial data, including land use and population density, are integrated within a grid-based model, with constraints such as ecological protection lines. In the park green space case, optimisation achieves 100% service coverage, reduces residents total travel time by 28.2%, increases 15-minute accessible population from 70.35% to 94.31%, and enhances efficiency. The Pareto optimal solution set illustrates critical trade-offs, while the optimised spatial layout demonstrates significant accessibility gains. This approach provides a robust decision-making tool for sustainable urban development, balancing ecological integrity, social equity, and economic viability in high-density environments. Keywords: multi-objective optimisation; MOO; non-dominated sorting genetic algorithm II; NSGA-II; sustainable landscape planning; genetic algorithm; GA. DOI: 10.1504/IJICT.2025.10074818
Abstract: Addressing the inability of traditional music teaching systems to dynamically adapt to learners personalised states, this study proposes deep reinforcement learning-MusicEdu a dynamic recommender system based on deep reinforcement learning. The framework constructs an intelligent agent that continuously perceives multidimensional learner states (skill proficiency, interests, fatigue) and dynamically optimises teaching-resource sequences via deep reinforcement learning (using proximal policy optimisation). This leverages a structured resource library derived from the Lakh Musical Instrument Digital Interface Dataset, annotated with metadata including difficulty, style, and technical attributes. Experimental validation across 20 weeks with five learner profiles demonstrates that deep reinforcement learning-MusicEdu significantly outperforms baselines, improving skill growth rate by 19.2% (p < 0.01) and user retention by 18.1%. The system enables personalised adaptive learning pathways, establishing an innovative decision making framework for intelligent music education. Keywords: deep reinforcement learning; DRL; music education; personalised recommendations; Lakh MIDI Dataset; adaptive learning. DOI: 10.1504/IJICT.2025.10074819
Abstract: Against the backdrop of increasingly fierce competition in the digital media industry, how to accurately predict operational effects has become the key to enhancing the competitiveness of media. Aiming at the problem of fusion redundancy caused by the existing research ignoring the mutual influence among cross-modalities, this paper first uses BERT and the improved visual transformer model to extract text and image features respectively. Then, cross-modal shared computing is utilised to enhance the complementarity among the features of each modal. Introduce text gating enhancement and use text information as prior knowledge to guide and improve the representation of image characteristics. Eventually, the fused characteristics are input into the classification layer for prediction. Experimental outcome indicates that the prediction accuracy rate of the suggested approach is 95.3%, which is at least 2.2% higher, significantly improving the accuracy of predicting the operation effect of digital media. Keywords: digital media; operation effect prediction; sentiment analysis; convolutional neural network; vision transformer. DOI: 10.1504/IJICT.2025.10074866
Abstract: As the field of educational assessment is growing, traditional ways of testing English reading ability cannot adequately show all the different kinds of information that learners use when they read. Because of this, how to employ multimodal learning behaviour data to make more accurate assessments is a popular topic in educational research right now. This research suggests the MLB-ERAM model for assessing English reading proficiency based on facts on how people learn in different ways. MLB-ERAM uses a lot of multimodal learning behaviour data and deep learning (DL) technology to get a whole picture of how well students can read. The experimental results reveal that the MLB-ERAM model works well with multimodal data, gets around the problems with standard assessment methods, and is a useful guide for the future growth of educational assessment technology. Keywords: multimodal data; learning behaviour; English reading proficiency assessment; DL. DOI: 10.1504/IJICT.2025.10074867
Abstract: Traditional manual arrangement struggles to efficiently handle complex backgrounds, frequent movements and variable lighting in dance videos; high-precision automatic techniques are urgently needed for digital analysis and protection. In this paper, a large-scale dance video dataset covering multi-ethnic, multi-scene and multi-illumination conditions is constructed, and the preliminary extraction of foreground region is realised by using the motion detection module combining frame difference method and optical flow method. Then, based on the improved U-Net structure, multi-scale feature fusion and attention mechanism are designed to enhance the segmentation ability of clothing and limb details, and the joint loss of Dice and cross entropy is used to improve the boundary accuracy. Experimental results show that the proposed method is better than U-Net and DeepLabv3+ in terms of IoU, Dice, precision, recall, F1-score, etc., and shows stronger robustness and near real-time processing speed in complex scenes. Keywords: intangible cultural heritage; ethnic dance; video image processing; dynamic segmentation algorithm. DOI: 10.1504/IJICT.2025.10074932
Abstract: This paper presents a digital twin (DT) framework aimed at predicting and managing injury risks in competitive sports. The proposed framework integrates biomechanical data, machine learning, and real-time monitoring to support athlete health and performance. DT technology enables detailed performance analysis and illness prediction through virtual athlete models. While previous studies have used wearable sensors and machine learning, they often lack cross-domain integration. This research addresses that gap by introducing a nine-stage DT development pipeline incorporating biomechanical data, PCA-based preprocessing, and musculoskeletal modelling. Validation through cross-validation and evolutionary data scenarios demonstrated the models robustness. XG Boost achieved the highest injury prediction accuracy (78%), with key predictors including hamstring force and muscle stiffness. Biomechanical simulations revealed stress patterns consistent with physiological behaviour, supporting clinical relevance. Comprehensive decision-support systems remain scarce. This work contributes toward safer, more personalised sports environments using a holistic, data-driven DT approach. Keywords: digital twin; DT; injury prediction; biomechanics; sports analytics; machine learning; musculoskeletal modelling; athlete monitoring; explainable AI; XAI; data-driven sports management; injury risk assessment. DOI: 10.1504/IJICT.2025.10074940
Abstract: Extreme weather is intensifying and demands forecasts that capture how hazards emerge and travel. To address non-stationary dependencies and rare tails, this paper introduces an adaptive spatiotemporal graph framework. In the scheme, first a fused relational graph blends physical priors with a learned adjacency to follow changing pathways; then a multi-scale temporal encoder links weak precursors to rapid surges; finally calibrated decision layers set cost-aware thresholds and deliver coherent maps with an audit trail. Experiments show gains at six-hour lead: area under the precision-recall curve rises by 8.5% over graph WaveNet and 22% over LSTM, fractions skill at ten kilometres improves from 0.44 to 0.52, the Brier score drops from 0.134 to 0.117, and false alarms fall by about ten percent, while seasonal scores remain between 0.41 and 0.53. Keywords: adaptive graph convolution; extreme weather; spatiotemporal dependence; teleconnections; early warning. DOI: 10.1504/IJICT.2025.10074941
Abstract: The explosion of internet news data has led to frequent information overload, which in turn has caused the dual problems of low recommendation quality and poor user experience. The research combines machine learning with image matching algorithms to provide users with interesting news content. Firstly, the news recommendation algorithm based on multi-task learning (MTL) constructs the input sequence through the users click history, maps the inputs of different tasks to the shared space and extracts text features. The image matching algorithm is adopted to record users historical preferences and capture the changing trends of interests. Experimental data show that the accuracy rate of the content-based recommendation (CB) algorithm reaches 78.27%. The collaborative filtering (CF) recommendation algorithm reached 87.97%. The recommendation accuracy of the joint recommendation algorithm all exceeds 90%. Moreover, the accuracy of the joint recommendation algorithm significantly exceeds that of the two traditional recommendation methods, CB and CF. Keywords: machine learning; image matching algorithm; collaborative filtering; news recommendations; content-based recommendation. DOI: 10.1504/IJICT.2025.10074942
Abstract: This research presents a knowledge graph-driven recommendation system for entrepreneurship courses, incorporating data collection, pre-processing, entity recognition, graph analysis, and recommendation algorithms. By combining semantic relationships with collaborative filtering, the system enhances personalisation, improving both accuracy and user satisfaction. Knowledge graphs provide a structured representation of entities and their relationships, enabling more relevant and context-aware recommendations. Leveraging this capability, the system aligns course suggestions with learners preferences and educational goals. Prior studies highlight the effectiveness of knowledge graphs in domains such as tourism, education, and e-commerce for improving recommendation precision. The proposed workflow follows sequential stages, including data pre-processing, knowledge extraction, graph construction, analysis, and algorithm development. integrating top-down ontology design with bottom-up entity extraction, guides knowledge graph creation. Experiments on a dataset of over 6,000 educational resources achieved stable accuracy after 80 training epochs using the BERT-BiGRU-MHSA-CRF framework. User evaluations confirmed higher engagement and satisfaction compared to baseline models. Keywords: knowledge graph; recommendation system; entrepreneurship education; data pre-processing; semantic analysis; collaborative filtering. DOI: 10.1504/IJICT.2025.10074943
Abstract: As semiconductor technology enters the nanoscale era, power control has become a major challenge impeding chip performance. Traditional heuristic methods fail to handle complex dynamic workloads, while pure reinforcement learning lacks stability and safety. This paper proposes a hybrid intelligent control framework integrating reinforcement learning with dynamic physical modelling. Offline-trained power and performance models offer prior guidance and safety constraints for online decisions. Experiments using Google production cluster data show the framework achieves an average power of 101.3 W 21.3% lower than on-demand strategies and a tail latency of 34.1 ms with only 1.2% violation rate. The energy efficiency reaches 15.6 instructions per joule, outperforming existing methods. This study provides an effective solution for energy-aware chip-level power management and introduces new ideas for intelligent cyber-physical systems. Keywords: power consumption control; reinforcement learning; dynamic modelling; energy efficiency optimisation; chip management. DOI: 10.1504/IJICT.2025.10074944
Abstract: emantic event analysis in sports videos faces challenges such as complex actions and high annotation costs. To address these issues, this paper proposes a novel framework that integrates domain knowledge with deep features. The approach first translates sports rules into computable spatio-temporal constraints, then designs a knowledge-injection network to guide deep models toward semantically critical regions. Finally, a knowledge-conditioned attention mechanism is introduced to fuse domain knowledge with visual features effectively. Experimental results on the SoccerNet dataset demonstrate that the proposed method achieves a mean average precision of 71.5%, outperforming strong baselines such as inflated 3D ConvNet and soccer background matting network by 13.3% and 3.6%, respectively. The framework shows significant improvements in detecting complex and sparse events, offering enhanced accuracy, robustness and generalisation capability with reduced reliance on large-scale annotated data. Keywords: semantic event analysis; domain knowledge; deep features; video understanding; sports videos. DOI: 10.1504/IJICT.2025.10074945
Abstract: Aiming at the challenges of spatio-temporal nonlinearity and physical consistency in regional wind speed prediction, this paper proposes graph attention network with physical constraints, a new model that fuses the graph attention network and meteorological equations. The model dynamically captures the complex relationship between meteorological stations through the graph attention mechanism and jointly optimises the loss function with the horizontal momentum equation as the physical constraint. Experiments based on high-resolution data in the Beijing-Tianjin-Hebei region from 20182021 show that the root mean square error and mean absolute error of graph attention network with physical constraints in 24-hour forecasts are 1.52 m/s and 1.11 m/s, respectively, which are reduced by 11.1% and 11.9% compared to the optimal baseline, and the R2 is improved to 0.948. Its excellent performance in extreme events provides a new paradigm for high-precision, interpretable weather prediction. Keywords: wind speed prediction; graph attention networks; physically informed machine learning; coupled meteorological equations; spatio-temporal prediction. DOI: 10.1504/IJICT.2025.10074946
Abstract: Short-term electricity load forecasting is a critical component of power system dispatch operations. As a core variable influencing load fluctuations, the precise quantification of meteorological factors impact has long been a research challenge. This paper proposes an innovative forecasting method integrating a diffusion model with a causal attention mechanism. This approach utilises the diffusion model to capture the randomness and uncertainty inherent in meteorological factors, while explicitly modelling the causal relationship between weather variables and electricity load through the causal attention mechanism. Experiments on public datasets demonstrate that the proposed method reduces prediction errors by 12% compared to traditional long short-term memory models, achieving over 90% prediction accuracy during extreme weather events. This provides a new pathway for refining the quantification of meteorological impacts and offers significant reference value for power system dispatch decision-making. Keywords: diffusion model; causal attention mechanism; quantification of meteorological impacts; extreme weather events. DOI: 10.1504/IJICT.2025.10074947
Abstract: This study addresses the efficiency limitations of traditional economic data analysis methods when processing large-scale, multi-source datasets under regulatory constraints. A blockchain-based system is proposed, featuring a novel sharding consensus algorithm with security and performance tradeoffs (SPTSCA) as its core component. Compared with existing approaches such as practical Byzantine fault tolerance (PBFT) and OmniLedger, the key innovations of SPTSCA include a dynamic shard adjustment mechanism for improved load balancing and an optimised consensus process that minimises communication rounds. Experimental results demonstrate that, due to more balanced shard formation, SPTSCA achieves up to a 1.49% increase in throughput compared with OmniLedger. More importantly, its performance significantly surpasses that of PBFT, with maximum throughput improvements of 143.1% and latency reductions of 89.1%. The algorithm enables secure large-scale economic data sharing, providing robust technical support for regulatory authorities. Keywords: legal regulation; economic data analysis; management system; blockchain. DOI: 10.1504/IJICT.2025.10074948
Abstract: Addressing the challenges of subjectivity and delayed feedback in English writing sentiment analysis, particularly for non-native academic contexts, this study proposes an automated framework based on a hybrid bidirectional encoder representations from transformers-long short-term memory model. The model integrates Berts contextual encoding capabilities with long short-term memorys sequential modelling strengths, leveraging attention mechanisms for three-dimensional granular sentiment analysis (tendency-intensity-object). Evaluated on a 20202023 subset of the international English language testing system writing dataset, the model achieves a sentiment classification accuracy of 91.5% (F1-score 0.89), outperforming baseline models by 12.3%. Educational applicability testing showed an 87.2% teacher approval for reducing feedback workload, confirming its efficacy in supporting writing pedagogy decisions. Key innovations include a multi-task with domain adaptation and a visualised feedback system, establishing a new paradigm for intelligent educational tool with real-time intervention capabilities. Keywords: sentiment analysis; english writing assessment; BERT-LSTM hybrid model; attention mechanism; educational feedback system. DOI: 10.1504/IJICT.2025.10074949
Abstract: Ice accumulation on insulators can lead to electrical breakdown, equipment damage, and line outages, making timely and accurate detection essential for maintaining the safe and stable operation of power systems. This paper proposes an ice accretion detection method for insulators based on You Only Look Once version 11 (YOLOv11), integrating image processing and deep learning techniques to achieve automated detection. A self-built dataset was used to fine-tune YOLOv11, enhancing the models accuracy and robustness in complex environments. Compared to its predecessors, YOLOv11 features an improved backbone network for more efficient feature extraction, advanced attention mechanisms for enhanced focus on critical regions, and an anchor-free detection head that reduces computational complexity while maintaining high precision. Multi-scale feature fusion ensures the accurate detection of ice accretion of various sizes, while dynamic label assignment optimises alignment between predictions and ground truth. Experimental results demonstrate that the fine-tuned YOLOv11-based algorithm achieves high mean average precision (mAP) and F1-scores on the test set, indicating robust detection performance. The proposed method not only enhances detection efficiency but also reduces labour costs, making it well-suited for large-scale power line monitoring. Keywords: ice accumulation; insulator icing detection; YOLOv11; ice accretion detection.
Abstract: To address the usability challenges faced by elderly users when operating smart home systems in the context of an aging population, this study proposes a smart aging-friendly home control algorithm framework based on multimodal interaction. The core of this framework lies in the innovative design of three key algorithms: a multimodal fusion decision-making algorithm that integrates speech recognition, simple gesture understanding, and touches intent analysis; an aging-friendly interaction optimisation algorithm; and a context-aware intelligent assistance algorithm. The proposed algorithms were validated through user simulation and comparative experiments. The results indicate that the algorithm framework effectively improves elderly users operational efficiency and task completion rates while significantly reducing cognitive load and operational error rates. This study provides core algorithmic support and practical design guidelines for constructing truly elderly-friendly smart home interaction systems. Keywords: multimodal interaction; age-friendly home; context awareness; smart home.
Abstract: Artificial intelligence (AI) and internet of things (IoT) convergence brings immense opportunity to convert the laboratory environment into intelligent, adaptive systems. This study proposes an integrated AI-IoT framework for smart laboratory engineering construction and engineering management team optimisation, which overcomes the current shortcomings in resource efficiency, task scheduling, and environmental control to some extent. In this system, real-time IoT sensor networks monitor ecological and operational conditions; meanwhile, LSTM models are applied for predictive environmental control, genetic algorithms for dynamic task scheduling, and SVM classifiers for human activity recognition. The framework was deployed in a research laboratory for six months, and the system achieved substantial improvements: energy consumption was reduced by 28.48%, equipment downtime by 54.37%, and task overlap and average task duration were significantly minimised. Additionally, predictive maintenance accuracy reached approximately 93.2%, eliminating passive interventions and improving equipment availability. Since intelligent task allocation incorporates fault tolerance considerations, workload imbalance in task execution is alleviated, and staff satisfaction is enhanced. Our results demonstrate that a collaborative AI-IoT approach can effectively improve infrastructure efficiency and worker productivity. In this context, the proposed framework provides a scalable, sustainable, and context-aware solution for next-generation laboratory environments in academic and industrial domains. Keywords: artificial intelligence; AI; internet of things; IoT; smart laboratory; engineering management; predictive maintenance; task scheduling; environmental monitoring.
Abstract: Temperature compensation is crucial for improving sensor accuracy and stability in high-precision measurement. Although radial basis function (RBF) neural networks perform well in nonlinear modelling, they face slow convergence, long training time, and limited accuracy. To address these issues, this paper proposes an improved RBF algorithm (QOLS-RBF) by combining quantum controlled-NOT (C-NOT) gates with orthogonal least squares (OLS) theory. The method quantises input data and applies quantum superposition, entanglement, and interference to enhance feature extraction and centre aggregation. It further integrates OLS screening with the maximum error compression ratio, using C-NOT gate evolution to reduce hidden layer nodes and accelerate convergence. Experiments with 85 training and 170 testing sensor datasets show that QOLS-RBF outperforms RBF, OLS-RBF, K-means RBF, and FCM-RBF in convergence speed, training time, error accuracy, and network compactness. This approach enables efficient temperature compensation and offers a promising tool for modelling complex nonlinear systems. Keywords: neural network algorithm; orthogonal least squares; OLS; sensors.
Abstract: For the insufficiency of traditional systems in automated data processing and predictive analysis capability, this study explored a visual data analysis system based on advanced artificial intelligence technology, integrating the three core functions of automated data preparation, intelligent recommendation and predictive analysis. Data cleaning was carried out by weighted k-nearest neighbours imputation and isolation forest algorithm. The unstructured data was handled utilising bidirectional encoder representations from transformers (BERT) models, and key patterns, trends and anomalies were discovered by means of association rule learning techniques. Relying on the autoregressive integrated moving average (ARIMA) model, the time series data was precisely forecasted. Distributed deployment supports the hardware and solves the system layout problem. The evaluation outcomes demonstrated that the ARIMA model performed the best in data prediction with an average prediction time of only 1.075 seconds, the lowest RMSE (7.19) and MAE (4.70), and the highest prediction accuracy (96.00%). This paper provides efficient and intelligent data support and solutions to help decision-making and strategic planning in various industries. Keywords: visual data analysis system; artificial intelligence technology; distributed deployment; predictive analysis; ARIMA model; data automation processing. DOI: 10.1504/IJICT.2025.10074751
Abstract: In order to improve the accuracy of French pronunciation correction, this study develops a multimodal feedback system, which adopts the improved wav2vec2 model to integrate the physiological features of articulation, and combines time-frequency analysis to extract the acoustic parameters. The developed system generates the targeted training materials through the dynamic knowledge graph and integrates the articulatory organ visualisation module. The selective spectrum enhancement strategy is designed to assist in the listening discrimination training. Experiments show that the feedback delay of the system is ≤ 155 ms, and the VOT recognition error is reduced by 9.2%; after ten weeks of training, the confusion rate of articulatory parts is reduced by 5.1%, and the accuracy rate of question rhymes reaches 79.2%. The results confirm that moderate multimodal feedback has a progressive optimisation effect on French pronunciation. Keywords: multimodal feedback systems; French language; acoustics; dynamic knowledge mapping. DOI: 10.1504/IJICT.2025.10074805
Abstract: With the rapid development of sports and computer technology, accurate sports movement analysis has become crucial for enhancing athlete performance and rehabilitation. Traditional methods face challenges such as difficulty in recognising multi-scene actions and inconsistent sequence lengths. To address this, a novel approach combining an early fusion network with human key point data is proposed. By integrating skeleton node information and using Neville interpolation, the method enhances feature extraction and temporal localisation. Experimental results show significant improvements: compared to traditional models such as LSTM and ST-GCN, the EF-GCN model proposed in this study achieves an increase in classification accuracy of up to 18.5% across various neural networks, and performance metrics such as accuracy, precision, recall, and F1-score improve by around 10%. This approach offers substantial advancements in motion analysis and holds great potential for future sports training and rehabilitation applications. Keywords: early fusion network structure; key points of the human body; sports movement analysis; Neville interpolation method; temporal positioning. DOI: 10.1504/IJICT.2025.10074850
Abstract: To address Japanese pronunciation error detection, this paper proposes a fusion method based on cross-modal attention mechanisms and constructs a Japanese pronunciation corpus. The model integrates audio Mel-spectrogram and visual lip-motion features through attention mechanisms, effectively capturing fine-grained cross-modal interactions and enabling precise phoneme-level error recognition. Evaluated on both the public corpus from Saruwatari Lab, University of Tokyo and a self-built corpus, the proposed approach achieves an accuracy of 92.3%, which is 3.1% higher than the best baseline model. Moreover, it maintains a robust accuracy of 85.3% under a low signal-to-noise ratio of 5 db, representing a 6.6% improvement compared to other methods. This study provides an effective and noise-robust tool for multimodal speech learning with strong potential for educational applications. The released corpus contains 50 hours of multimodal data with detailed annotations, offering comprehensive support for Japanese language teaching and advanced speech technology development. Keywords: cross-modal learning; pronunciation error detection; Japanese speech processing; attention mechanisms; corpus construction. DOI: 10.1504/IJICT.2025.10074807
Abstract: Oral fluency is a key indicator for evaluating the professional skills of broadcast hosting. To address the current research gap in modelling deep semantic associations for spoken fluency, this paper first utilises Res2Net for multiscale feature extraction from broadcast hosts' speech. Subsequently, a pause prediction module is proposed. This module predicts multiple types of pause labels based on the original text. It then predicts a Gaussian mixture distribution for each phoneme and achieves diverse phoneme durations through random sampling. Finally, an autoregressive large language model and a discriminative module based on transformer are proposed. This module is applied at each time step of the autoregressive process and prevents misalignment phenomena via the transformer and judging mechanism. Experimental results show that the proposed model achieves an evaluation accuracy of 93.35% and a word error rate of 0.7%, enabling high-accuracy fluency evaluation for oral speech. Keywords: spoken fluency assessment; feature extraction; Res2Net model; autoregressive large language model; transformer model. DOI: 10.1504/IJICT.2025.10074862
Abstract: Hospital social security settlement is a core link connecting medical services, medical insurance systems and patient interests, with its operational efficiency directly affecting medical service quality and social security system sustainability. Expanded medical insurance coverage, surging daily settlements and frequent policy adjustments make traditional static scheduling unable to adapt to system dynamics, causing long patient waits, low terminal utilisation and high verification failures. This study proposes a dynamic scheduling framework for hospital social security settlement based on multi-agent reinforcement learning, with four intelligent agents for distributed decision-making, plus a multi-objective reward function and constrained action mechanism. Experiments with real data from a tertiary Grade A hospital show the framework cuts average settlement delay by 38.2% and 21.5%, raises terminal utilisation by 27.6% and maintains over 99.5% compliance. It offers an intelligent solution to boost settlement efficiency and supports medical insurance service digital transformation. Keywords: hospital social security settlement; dynamic scheduling; multi-agent reinforcement learning; MARL; intelligent agent; settlement efficiency. DOI: 10.1504/IJICT.2025.10074806 |
Open Access
Keep up-to-date
-
Our Blog -
Join us on Bluesky -
Follow us on X -
Visit us on Facebook - Our Newsletter (subscribe for free)
- RSS Feeds
- New issue alerts
- Inderscience is a member of publishing organisations including:
-
-
