International Journal of Intelligent Systems Design and Computing (4 papers in press)

Regular Issues

• The conceptual frameworks for global acceptance of artificial intelligence in the post-pandemic era  
  by Efosa Carroll Idemudia 
  Abstract: In 2024, the generative AI market was valued at $128 billion. To date, numerous firms, organisations, governments, and institutions have invested millions and billions of dollars in artificial intelligence to make informed decisions and gain a competitive advantage. To provide a holistic view and insights on how companies and institutions can utilise artificial intelligence to solve complex real-world problems, make informed decisions, and gain a competitive advantage in the post-pandemic era, we conducted our study. The theoretical background for our models and frameworks is the stakeholder theory. Our models and frameworks provide insights and understanding into how firms, companies, organisations, governments, and institutions can adapt to and adopt a wide range of artificial intelligence platforms. Our frameworks identify the key variables that influence the acceptance and usage of artificial intelligence. Our study has many managerial and research implications.
  Keywords: stakeholder theory; post-pandemic era; data; artificial intelligence; diverse team; algorithm design; public perception; environmental factors.
  DOI: 10.1504/IJISDC.2026.10077585
   
  
• A survey of textual information extraction methods for knowledge graph construction  
  by Rui Chen, Hang Li 
  Abstract: Knowledge graphs play an increasingly important role in cognitive intelligence applications, making efficient and accurate extraction of structured knowledge from unstructured text a core challenge in natural language processing. This paper presents a systematic review of textual information extraction methods for knowledge graph construction, covering key subtasks such as named entity recognition, relation extraction, attribute extraction, and event extraction. We review the evolution of extraction techniques from rule-based and statistical approaches to deep learning models and pretrained language model-based paradigms. Particular attention is given to joint extraction methods designed to alleviate error propagation in traditional pipeline architectures. Furthermore, the survey discusses emerging extraction frameworks based on large language models, highlighting the potential of prompt-based and in-context learning for more flexible and generalised knowledge extraction. Finally, we summarise representative application scenarios and analyse major challenges, including low-resource adaptability, complex semantic modelling, and dynamic knowledge graph construction.
  Keywords: knowledge graph construction; information extraction; named entity recognition; NER; relation extraction; pretrained language models; PLMs.
  DOI: 10.1504/IJISDC.2026.10078923
   
  
• A fusion-hybrid recommender system using GRU, graph attention networks, and product feature encoding for e-commerce  
  by S. Sangeetha, Anjana Devi Saravanan, Nithya Shrree Saravanan Monimekalai,, Kathiresh Durairaj, Saisree Anusha Balaji, Niranjana Ramesh 
  Abstract: Recommender systems are a core component of modern e-commerce platforms, yet building models that jointly exploit user behaviour, product-to-product relationships, and rich product metadata remains challenging. In this work, we propose a fusion-hybrid recommender system (FHRS) that integrates three complementary representation spaces: sequential user encoder, a product graph encoder built with graph attention networks (GAT), a high-dimensional content feature encoder using TF-IDF and dense projections of product metadata. These three views are fused at the representation level and trained end-to-end on an Amazon All Beauty subset. On top of the fused backbone, we implement four recommendation strategies: user-based (personalised), item-graph (similar products), product-feature similarity, and a hybrid method that combines user preferences with item similarity. Experiments show that our fused model achieves a hit rate and Recall@10 of 97.94%, with NDCG@10 of 0.9518 and MRR of 0.9428 on the test set. In contrast, a baseline implementation — developed separately using only GRU + GAT without product feature encoding and with fusion applied only at inference — achieves 78% accuracy. Our results demonstrate that: 1) learning a fused representation during training; 2) integrating pre-processed metadata features, significantly improves recommendation quality compared with graph-only or sequence-only approaches and inference-level hybrids.
  Keywords: GRU; graph attention networks; GATs; hybrid recommendation; e-commerce; representation fusion.
  DOI: 10.1504/IJISDC.2026.10081608
   
  
• The analysis of the use of artificial intelligence for research, innovation, and competitive advantages in the post-pandemic era  
  by Efosa Carroll Idemudia, Daniel Owunwanne 
  Abstract: To date, artificial intelligence (AI) is significantly impacting all disciplines, institutions, firms, and organisations. Our study provides a comprehensive review of companies, institutions, firms, organisations, and educational institutions that are adopting a wide range of AI platforms and domains. AI is used in writing and research, personalised learning, creative content generation, cybersecurity defence, communication, collaboration, project management, decision support, educational settings, market research, academic integrity, task automation, adaptive learning, meeting summaries, software development, skill development, etc. Our research has significant implications for both research and management applications. In addition, our study provides top management with insights and guidance on specific strategies for implementing effective and efficient AI policies.
  Keywords: artificial intelligence; AI; personalised learning; cybersecurity defence; collaboration; decision support; educational settings; market research.
  DOI: 10.1504/IJISDC.2026.10081787