International Journal of Information Technology and Management (12 papers in press)

Regular Issues

• Open Access

  Research on the Safety Monitoring of the Entire Process of Green and Low-Carbon Expressway Construction Based on the Improved Apriori Algorithm ( Free Full-text Access )

  by Yongjian Guo, Chenglin Xing, Xibo Wang, Yali Liang 
  Abstract: To overcome the problems of high false acceptance rates, high false rejection rates, and long response times in current monitoring methods, a new safety monitoring method for the whole process of green and low-carbon expressway construction based on an improved Apriori algorithm is proposed. Data on green and low-carbon expressway construction are collected, and an improved isolated forest algorithm is used for outlier detection and data preprocessing. An improved Apriori algorithm is designed, which first divides regions and then integrates them. This algorithm uses the processed data as a basis to assess construction safety status and achieve safety monitoring throughout the entire construction process. Experimental results show that the false acceptance rate of the proposed method remains stable between 4.12% and 7.95%, the false rejection rate remains stable between 3.84% and 7.28%, and the response time remains stable between 0.25s and 0.37s.
  Keywords: Improved Apriori algorithm; Expressway; Green and low-carbon construction; The Whole process; Safety monitoring; Improved isolated forest algorithm.
  DOI: 10.1504/IJITM.2026.10080128
   
  

• Open Access

  Automatic Identification Systems for Accounting Information Disclosure Based on Deep Learning ( Free Full-text Access )

  by Yingjian Su, Chengyun Zhou, Hongge Yu 
  Abstract: The large-scale and complex structure of corporate accounting disclosure data makes it difficult to achieve accurate automatic identification and information classification. This study proposes a deep learning-based identification framework using the pre-trained BERT model to improve disclosure classification performance. Public data were collected from the Soochow Securities Research Institute and preprocessed through data cleaning, missing value interpolation, and Z-score normalization. A 10-fold cross-validation strategy was adopted, and Token Embedding, positional encoding, and Transformer Encoder were constructed to extract semantic features from accounting information. Additional pooling and classification layers were introduced after the Transformer Encoder, with the Masked Language Model used as the training objective. Experimental results demonstrate that the proposed BERT model achieves an identification accuracy of 97.1% and an average recall of 0.985, indicating its effectiveness in automatic accounting disclosure classification.
  Keywords: Deep Learning; Accounting Disclosure Information; Automatic Identification System; Classification and Identification Ability.
  DOI: 10.1504/IJITM.2026.10081178
   
  

• Open Access

  Customer Churn Prediction for B2C E-Commerce using K-Means and Support Vector Machine Model ( Free Full-text Access )

  by Jian Gao 
  Abstract: E-commerce faces high customer churn, making loss prediction vital for retention. This study proposes a method combining k-means customer segmentation with SVM and logistic regression (LR) prediction, using B2C purchasing behaviour data. Customers were clustered by activity continuity: those with declining engagement and no transactions were labelled churn, others non-churn. After segmentation, prediction performance improved significantly. The SVM model outperformed LR, achieving average accuracy of 0.9158, recall of 0.9723, precision of 0.8569, and AUC of 0.9800, versus LRs 0.9062, 0.9497, 0.8534, and 0.9601 respectively. The results confirm that integrating clustering enhances churn prediction and provides valuable insights for B2C e-commerce relationship management
  Keywords: E-commerce; k-means; SVM model; customer churn prediction.
  DOI: 10.1504/IJITM.2026.10081179
   
  

• Open Access

  Blockchain and IoT-Driven Dynamic Enterprise Cost Management: A TKG-GAT Intelligent Inference Approach for Supply Chain Collaboration ( Free Full-text Access )

  by Limei Yin, Chengyong Liu, Jiayao Lin 
  Abstract: Traditional enterprise cost management suffers from delayed data acquisition, limited cross-entity sharing, and weak dynamic response. This paper constructs a dynamic cost management model driven by blockchain and IoT, enabling real-time production status sensing, trusted cost recording, and cross-organizational collaboration. An intelligent inference module combining Temporal Knowledge Graph Embedding and Graph Attention Network (TKG-GAT) is developed to model dynamic entity relationships, predict cost trends, identify key drivers, and evaluate alternative business schemes. Experimental results show that TKG-GAT achieves a PICP of 95.0%, PINAW of 0.353, HR@3 of 0.895, and an average comprehensive decision support score of 0.905. The model effectively transforms cost management from ex-post accounting to proactive, real-time decision support, providing a quantitative foundation for budget adjustment, supply chain optimization, and production planning in digital environments.
  Keywords: Enterprise Cost Management; Blockchain Perspective; Internet of Things; Time-Series Knowledge Graph; Dynamic Cost Inference.
  DOI: 10.1504/IJITM.2026.10081313
   
  

• Open Access

  An Empirical Study of the Relationship Between Financial Development Foreign Trade and Energy Consumption Based on the ARDL Model ( Free Full-text Access )

  by Jiahui Jin, Xinrui Jing 
  Abstract: This study uses a hybrid HCSARDL-GLSM estimator and Granger causality test, we investigate the long-run and causal relationships among these variables. Empirical results show significant cointegration: economic growth and energy consumption significantly increase consumption-based CO emissions, while financial development and foreign trade significantly reduce them. Causality tests confirm significant causal linkages from economic growth, energy use, financial development and foreign trade to CO emissions. The proposed method outperforms traditional approaches in robustness and estimation accuracy. Quantitatively, the long-run coefficient of financial development on consumption-based CO emissions is -1.78 at the 4% significance level, and foreign trade yields a long-run negative coefficient of -0.945 at the 3% significance level; the HCSARDL-GLSM model achieves an optimal fitting R? of 0.947 with a minimum MSE of 0.021, far superior to conventional ARDL-based estimators.
  Keywords: Economic growth; financial development; foreign trade; Cross-Sectional Autoregressive Distributed Lag.
  DOI: 10.1504/IJITM.2026.10081348
   
  

• Open Access

  Resource Optimisation Allocation and Policy Evaluation Based on Cost-Benefit Analysis ( Free Full-text Access )

  by Lili Wang, Yaozong Zhu, Jiayao Lin 
  Abstract: This article presents an organized dynamic coupling methodology aimed at attaining endogenous equilibrium via the iterative interaction of higher-level welfare optimization and lower-level agent behavior. Analyzing the interaction between if a social cost-benefit model derived from environmental externalities and an agent-based production response model, significant parameters such as price susceptibility and policy expectations were determined from historical events. A bi-level program was created to improve macro allocation and measure policy consequences which was coupled with a region-based compensation system. The analysis showed that the preceding extent of resource mismatch loss in the traditional coal sector decreased to 18.7%, while the regional gini decreased to 0.3 or less.
  Keywords: Cost-Benefit Analysis; Optimal Resource Allocation; Policy Evaluation; Agent Modeling; Two-Level Programming.
  DOI: 10.1504/IJITM.2026.10081349
   
  

• Open Access

  Precise Perception of Enterprise Financial Risk based on Joint Improved Random Forest Algorithm ( Free Full-text Access )

  by Xuena Lin, Yan Xue 
  Abstract: Aiming at the shortcomings of traditional methods in prediction accuracy, risk recall rate, and computational efficiency, a precise perception method of enterprise financial risk through joint improved random forest algorithm is proposed. Firstly, establish a comprehensive financial risk assessment index system; secondly, the grey relational clustering algorithm is introduced to reduce the dimensionality of the original feature space and extract core feature clusters highly correlated with the risk state; finally, the beetle optimisation algorithm is used to adaptively global optimise key hyperparameters such as the number of decision trees and maximum depth in the random forest, in order to improve the accuracy and generalisation stability of the model. The experimental results show that this method has an accuracy rate of over 90% in financial risk perception, a recall rate of over 0.9, and a perception time of only 9.8 s when the data volume increases to 180 GB.
  Keywords: Improved random forest; Enterprise financial risk; Risk perception; Beetle optimization algorithm.
  DOI: 10.1504/IJITM.2026.10081400
   
  
• Can Blockchain Help Organizations Streamline Their Operations The Case of Luxury Brands  
  by Sarah Bouraga 
  Abstract: Luxury brands have been facing various pain points over the years, such as counterfeiting or the global issue of sustainability. They also have to create a deep connection with their consumers, making customer relationships essential. Blockchain can help address these issues and opportunities. Using a multiple-case study where we apply an exploratory and inductive approach, we address the following research questions: (i) How can blockchain streamline luxury brand business processes? (ii) How can luxury brands use blockchain to bring value to customers? (iii) How can DApps developers/designers make existing blockchain-based solutions more effective and efficient? This study allows us to draw various conclusions that can have implications in theory and practice In particular, the contribution lies in its identification of strategic and technological concerns around the implementation of a blockchain-based solution, the importance of sustainability, the recognition of characteristics of the blockchain-based application, and the acknowledgment of remaining
  Keywords: Blockchain; Non Fungible Token (NFT); Sustainability; Luxury Brand; Fashion; Counterfeiting .
  
  
• Can Blockchain Help Organisations Streamline Their Operations? The Case of Luxury Brands
  by Sarah Bouraga 
  Abstract: Luxury brands have been facing various pain points over the years, such as counterfeiting or the global issue of sustainability. They also have to create a deep connection with their consumers, making customer relationships essential. Blockchain can help address these issues and opportunities. Using a multiple-case study where we apply an exploratory and inductive approach, we address the following research questions: (i) How can blockchain streamline luxury brand business processes? (ii) How can luxury brands use blockchain to bring value to customers? (iii) How can DApps developers/designers make existing blockchain-based solutions more effective and efficient? This study allows us to draw various conclusions that can have implications in theory and practice In particular, the contribution lies in its identification of strategic and technological concerns around the implementation of a blockchain-based solution, the importance of sustainability, the recognition of characteristics of the blockchain-based application, and the acknowledgment of remaining challenges.
  Keywords: Blockchain; Non Fungible Token (NFT); Sustainability; Luxury Brand; Fashion; Counterfeiting .
  
  
• VRPM-HCM: Enhancing Cloud Power Efficiency Through Hybrid Machine Learning and Model Predictive Control  
  by Sai-Feng Zeng 
  Abstract: The proliferation of cloud computing has intensified energy consumption in data centres, demanding innovative solutions to balance power efficiency and performance. Traditional methods struggle to address dynamic interactions between virtual machines, workloads, and hardware in virtualised environments. This paper proposes VRPM-HCM, a novel framework that synergises model predictive control with machine learning to optimise power consumption while ensuring service-level agreement compliance. The framework introduces a dual-layer control mechanism for real-time CPU frequency adjustments and VM migrations. Experiments demonstrate that VRPM-HCM achieves 23.9% average power savings, reduces SLA violations by 1.2%2.7%, and maintains 91%94% resource utilisation, outperforming state-of-the-art baselines. These results validate its effectiveness in harmonising energy efficiency, performance guarantees, and hardware longevity in dynamic cloud environments.
  Keywords: virtual cluster; quality of service; power control; control theory.
  DOI: 10.1504/IJITM.2025.10076874
   
  
• Security Assessment and Improvement of Artificial Intelligence Technology in Information Retrieval  
  by Baolin Zheng, Shuqin Han, Zaihui Cao 
  Abstract: In view of dynamic security threats and malicious information in information retrieval, existing methods struggle to enhance intelligence and ensure security. This paper proposes an LSTM-DNN model integrating Long Short-Term Memory (LSTM) and Deep Neural Network (DNN). Retrieval data are collected and preprocessed, DNN extracts and fuses features, while LSTM captures temporal changes and identifies threats. The combined model strengthens security assessment and response to dynamic risks. Experiments show that with time-series feature capture and deep feature fusion, the model achieves precision and recall of 0.90 and 0.88 under imbalanced attack samples, demonstrating adaptability, robustness, and offering an effective security solution.
  Keywords: Information Retrieval System; Security Assessment; Dynamic Security Threats; LSTM-DNN Model; Malicious Information Identification.
  DOI: 10.1504/IJITM.2026.10080324
   
  

Special Issue on: OA Information Management and Information Visualisation - Part 3

• Open Access

  The role of business analysis in AI-driven digital transformation ( Free Full-text Access )

  by Korede J. Oluwamola  
  Abstract: Artificial intelligence (AI) is a key driver of digital transformation, enabling organizations to enhance decision-making, automate processes, and generate data-driven insights. However, realising its full potential requires alignment with organisational strategy and processes. This study examines the role of business analysis in facilitating AI-powered digital transformation through a qualitative systematic review based on the PRISMA framework. Twelve peer-reviewed studies across diverse organisational contexts were analysed to identify patterns of AI implementation. Findings indicate that business analysis mediates the translation of AI capabilities into strategic value by supporting process redesign, governance, data readiness, stakeholder engagement, and ethical practices. Rather than serving solely as a technical support function, business analysis emerges as a dynamic organisational capability integrating strategic, operational, and cultural dimensions. The study provides practical insights for organisations seeking responsible AI adoption to maximise long-term business value.
  Keywords: artificial intelligence; AI adoption; digital transformation; business analysis; organisational strategy; process redesign; governance; data readiness.
  DOI: 10.1504/IJITM.2026.10079741