Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
Investigating the Impact of AI-Driven Predictive Analytics on Hyper-Personalized Marketing in Niche Retail Markets
No full textThis manuscript presents a well-executed and timely investigation into how AI-driven predictive analytics can empower niche retail markets through hyper-personalized marketing. It addresses a significant gap in the current literature by focusing on smaller retail businesses that are often excluded from mainstream AI applications. By demonstrating both quantitative and qualitative outcomes from real-world implementations, the study offers valuable insights into the democratization of AI technology. The findings are relevant for both scholars and practitioners, particularly those interested in digital transformation, SME competitiveness, and customer-centric marketing strategies.
This research uses both qualitative and quantitative methods as a mixed-methods design. The research uses case studies of particular niche retailers who applied AI tools which include Google Analytics and Dynamic Yield as well as predictive modeling software to optimize their marketing operations. Survey responses along with performance measurement data from businesses participate in the study to assess AI-based marketing strategies\u27 success rate. The qualitative aspect gives background and perception into the issues and advantages of AI in real-life scenarios, whereas the quantitative aspect reveals trends in customer involvement, revisit rate, as well as income increase attributable to the use of AI.
The key findings demonstrate that AI-powered predictive analytics has a tremendous impact on the preciseness of customer targeting and increases sales by offering the personalized clients’ offers, product suggestions and timely messages. Businesses documented a 30% jump in customer retention rates while showing better results in email click-through and conversion rates. The research shows that businesses at any size can successfully implement these technologies through free or low-cost AI platforms which work well for their budget.
Finally, the study proves that AI can be an effective enabler for niche retailers who want to improve their marketing efforts regardless of the vast amounts of money invested. Subscribing to AI-assisted predictive analytics, small business can develop highly relevant and individualized experiences for their customers, thus nurturing loyalty and sustainable growth in the competitive markets
The Hidden Challenges of Building a Scalable Data Clean Room
No full textData clean rooms have emerged as critical infrastructure for enabling privacy-preserving collaborative analytics across heterogeneous data ecosystems. This article presents a comprehensive examination of the architectural and operational barriers organizations face when implementing scalable clean room solutions and offers field-tested architectural patterns and system-level strategies to overcome these challenges. Drawing on over three years of production-grade deployments, the study identifies key implementation bottlenecks related to data schema standardization, multi-cloud security integration, and resource-efficient privacy-preserving computation. The findings are applicable across various industry sectors and provide actionable insights to support secure data collaboration while ensuring regulatory compliance, cost efficiency, and operational scalability. This research addresses critical limitations of current clean room models by proposing concrete technical solutions for cross-cloud data collaboration architectures that accommodate diverse data volumes, complex privacy requirements, and evolving compliance frameworks
Machine Learning Systems for Predicting Consumer Behaviour
No full textThe development and implementation of machine learning systems for predicting consumer behaviour is a crucial stage in the digital transformation of business. Machine learning offers unique opportunities for analysing large volumes of data and identifying hidden patterns, significantly enhancing the decision-making process in marketing. The use of algorithms such as neural networks, decision trees, and clustering methods allows businesses to predict future consumer preferences and create personalised offers. This, in turn, increases customer loyalty and the effectiveness of marketing campaigns. The review discusses key approaches to applying machine learning in the context of consumer behaviour analysis, along with examples of successful implementation of these technologies in business practices. The findings revealed that decision trees serve as a tool for visualising decision-making processes, helping marketers segment audiences and predict customer behaviour. Classification methods allow users to be assigned to specific categories, which simplifies the development of personalised offers and advertising campaigns. Companies like Google and Facebook are actively developing deep learning-based tools, contributing to the further integration of AI into business processes. The ability to predict consumer behaviour with high accuracy is becoming an integral part of future marketing strategies, and deep learning is already significantly transforming approaches to product and service promotion. The high accuracy of predictions helps optimise supply chains and reduce operational risks, making machine learning a key element of modern marketing. The main findings of the study emphasise that the integration of machine learning systems enables businesses to manage inventory more efficiently, reduce risks, and adapt to changes in demand
AI-driven Detection and Prevention of Deepfakes in National Security
No full textThe rapid advancement of artificial intelligence has enabled the creation of highly realistic synthetic media, known as deepfakes, which pose significant threats to national security. This research explores the application of AI-powered tools to detect and mitigate deepfakes in defense, intelligence, and governmental communication channels. The primary objective of the study is to evaluate the effectiveness of current AI-driven detection techniques and propose robust mitigation strategies that can be integrated into national security frameworks.
A mixed-methods approach was employed, combining a comprehensive review of state-of-the-art detection algorithms—including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformer-based models—with qualitative analysis of their application in real-world security scenarios. Additionally, simulated deepfake attack scenarios were used to test detection accuracy, response time, and potential countermeasure efficacy.
The findings indicate that while AI-based detectors can achieve high accuracy under controlled conditions, their performance degrades with adversarially modified or low-quality content. Moreover, current systems lack seamless integration with national infrastructure, highlighting a critical gap in operational readiness. The study also identifies the importance of multi-layered defense systems incorporating forensic analysis, real-time monitoring, and public awareness initiatives.
In conclusion, while AI-powered tools offer promising capabilities in identifying and mitigating deepfakes, they must be supported by policy frameworks, inter-agency collaboration, and continuous technological advancement to be effective in safeguarding national security interests
A Review on RSA-Huffman Hybrid: towards Secure, Compressed and Adaptive Network Data Transmission
No full textThe RSA-Huffman hybrid model, which combines asymmetric encryption and lossless compression, has become a viable answer to the increasing demands for safe and effective data transfer. The review evaluates the existing RSA-Huffman techniques based on implementation by application domains (cloud, IoT, Mobile etc), the studies were selected and categories based on their implementation domain, performnces metrics of each and the year article was published. It asses their real-world feasibility, security-efficiency trade-offs, and techniques. Results show improved confidentiality (e.g., high NPCR scores) and compression ratios (50–90%), but they also highlight important drawbacks, including inconsistent benchmarking, fidelity degradation in lossy implementations, computational overhead (notable latency spikes), and new vulnerability risks (e.g., quantum threats). Future developments necessitate uniform evaluation procedures, quantum-resistant modifications, and lightweight optimizations to allow for scalable deployment in resource-constrained systems, even though they are successful in particular situations
Evaluating the Performance, Quality and Accessibility of Ghanaian Government Portals Using Diagnostic Tools
No full textIn today\u27s world, e-government has become an increasingly important aspect of public administration. With the widespread use of the internet, many governments have developed websites and portals to provide services and information to citizens. However, the success of these e-government programs depends on several critical factors, such as accessibility, quality, and efficiency. This research evaluated 20 e-government websites and portals in Ghana using a set of industry-standard Web diagnostic tools (TAW3 and Google PageSpeed Insights and W3C Web Validators). The results are presented in this paper. The findings show that none of the evaluated sites meet the recommended Priority AA conformance standard, and there are significant weaknesses in accessibility, quality, and performance. Most of the websites did not follow the Web Content Accessibility Guidelines (WCAG)\u27s robust, perceivable, operable, and understandable principles, which resulted in non-compliance problems like slow page loads and broken connections. These results underline the necessity of substantial enhancements to Ghana\u27s e-government websites\u27 usability, effectiveness, and efficiency to improve the citizen experience and guarantee the success of e-government initiatives
Comparative Insights on Centralized and Individual Models Using Snowflake and Google Big Query
No full textData sharing plays a vital role in today’s digital ecosystem, allowing businesses, governments, and individuals to exchange information on a massive scale. Cloud-native data platforms have significantly transformed traditional data management practices by introducing scale architectures, decoupled storage and compute models, cost-efficiency and resource governance. This paper investigates two distinct paradigms, (Centralized and Individual) data sharing models. Centralized architecture offers consolidated governance, multi-tenant scalability, and advanced analytics enablement, whereas individual data control underlines autonomy, privacy-preserving access, and federated governance. The comparative explores both paradigms through the capabilities of two leading platforms: Snowflake and Google Big Query (GBQ). This analysis highlights operations implications, governance complexity and data collaboration potentials. This paper also covers the pathways for organizations to adopt hybrid data sharing strategies with balance agility, regulatory compliance and efficiency in multi cloud environment
Improving Learning Quality through Digital Information Systems in Zambian Higher Education
No full textThe usage of digital information systems has significantly enhanced learning quality among students in higher learning institutions by promoting accessibility, interactivity, and efficiency in academic processes. These systems including Learning Management Systems (LMS), digital libraries, and data analytics platforms facilitate personalized learning experiences, streamline administrative tasks, and enable real-time feedback between lecturers and students. This study adopted a mixed-methods research design, combining both quantitative and qualitative approaches to gain a comprehensive understanding of how digital information systems are used to improve learning quality among students in Zambian higher learning institutions. The study was conducted in three higher learning institutions within Lusaka district, Zambia. The target population consisted of students and lecturers from selected higher learning institutions in Zambia with a sample size of 264. The data collection process involved distributing the questionnaires (quantitative data) and conducting individual interviews (qualitative data) to the selected respondents. The quantitative data collected through the questionnaires were analyzed using appropriate statistical methods, such as descriptive statistics using SPSS (statistical package for social sciences) and Microsoft excel whereas the qualitative data from semi structured interviews were analyzed thematically. The study found that the integration of digital information systems such as Learning Management Systems (LMS) and online assessment tools led to increased student participation, improved comprehension, and higher academic performance due to flexible access to learning materials and timely feedback. Additionally, the findings showed that students reported a positive perception of digital learning platforms, with many indicating increased motivation and satisfaction. Based on the findings, the study recommended that higher learning institutions should invest in continuous training and capacity building for both students and academic staff to effectively use digital information systems. Additionally, to sustainably improve learning quality through digital information systems in Zambian higher education, the government through the Ministry of Education and in collaboration with higher learning institutions should develop and implement a comprehensive National Digital Learning Policy Framework
Understanding AI Adoption in Higher Education: A Systematic Review of Technology Acceptance Model, Technology Readiness Index, and the Integrated Technology Readiness and Acceptance Model
No full textArtificial Intelligence (AI) is reshaping education through numerous pedagogical possibilities. To reap its benefits, it is critical to understand the factors affecting its adoption by users. Among various Information Systems (IS) theories applied in this context, the Technology Acceptance Model (TAM) and the Technology Readiness Index (TRI) are the most prominent. This systematic literature review (SLR) examines 25 studies, 20 based on TAM, 3 on TRI, and 2 combining both (TRAM or TRITAM) to explore how Artificial Intelligence (AI) technologies are being adopted in higher education. PRISMA guidelines were followed to ensure transparency and reproducibility of search results. Data extraction focused on key parameters like study design, analysis technique, technological focus, population studied, sample size used, theoretical lens applied, factors explored, and the key findings. The findings reveal that cognitive factors, Perceived Usefulness (PU), and Perceived Ease of Use (PEOU) are the most significant predictors of Behavioural Intention (BI) to adopt AI. Additionally, AI-specific constructs such as AI Literacy, AI Explainability, and Co-Creation Intention were also studied in a few studies. Actual technology use has been rarely measured, indicating a recurring intention–actual use gap. The review also highlights crucial research gaps, such as a lack of longitudinal studies, use of only self-reported measures, underrepresentation of inclusive education professionals and decision-makers, and absence of cross-cultural comparisons. Thus, by synthesizing core findings, this review offers practical guidance for designing inclusive, evidence-based AI adoption strategies in higher education
AI-Driven Work Order and Asset Management Systems in Facility Operations Using Natural Language Processing
No full textManual work orders and asset management systems often result in delays, inefficiencies, and communication errors in facility operations. This study proposes an AI-driven framework that employs Natural Language Processing (NLP) to automate the classification, prioritisation, and processing of maintenance requests, as well as the tracking of asset lifecycles. A fine-tuned BERT-based NLP model was developed to extract critical information, such as fault type, urgency level, and asset identifiers, from unstructured maintenance text logs. Integrated into a decision support module, the system automatically generates structured work orders and recommends technician assignments based on asset history and task severity. Evaluation using over 10,000 real-world maintenance logs showed that the model achieved 91% classification accuracy and reduced work order processing time by 45%. The findings underscore the potential of NLP to enhance the responsiveness, efficiency, and intelligence of Computerised Maintenance Management Systems (CMMS). This research contributes to the digital transformation of facility management by demonstrating the value of AI in enabling proactive and data-driven maintenance operations