Asian Journal of Research in Computer Science
Not a member yet
792 research outputs found
Sort by
The Transformative Role of Augmented Reality (AR) and Virtual Reality (VR) in E-commerce and Digital Marketing: Enhancing Consumer Engagement and Trust
No full textThis review examines the transformative role of Augmented Reality (AR) and Virtual Reality (VR) technologies in e-commerce and digital marketing. AR and VR provide immersive and interactive experiences, bridging the gap between traditional in-store and online shopping by enabling consumers to visualize and interact with products virtually. The integration of these technologies addresses critical challenges in online retail, such as product visualization, consumer trust, and engagement. AR facilitates real-time, contextual interactions, such as virtual try-ons and spatial placements, while VR creates fully immersive environments replicating or enhancing physical shopping experiences. The paper synthesizes current research, exploring the impact of AR and VR on consumer behavior, customer brand engagement, and purchase intentions. Additionally, it discusses the potential of these technologies to drive higher conversion rates, reduce return rates, and foster personalized shopping experiences. Despite their promise, challenges such as cost, accessibility, and technical limitations remain. This review highlights the need for strategic implementation and further research to maximize the benefits of AR and VR in reshaping the digital marketing and e-commerce landscape
Ensemble Learning Techniques for Rice Nutrient Disease Deficiency Detection and Prediction Analysis
No full textRice is a vital food source and its nutritional composition, including essential minerals and vitamins, significantly impacts human health. Understanding nutrient deficiencies and diseases in rice is crucial for promoting healthy and sustainable agriculture and preventing related health problems. Rice grain mostly suffers from production issues triggered by nutrient imbalances like potassium, phosphorus, and nitrogen. Generally, nutrient deficiencies in rice plants show stimulation due to differences in leaf colour. Leaf features provide nutrient shortage classification of colour and shape. This study presents ensemble learning to classify rice crop nutrient deficiencies. The datasets were taken from the Kaggle data source. It consists of hundreds of rice leaf images, it can be divided into different classes. They can represent deficiencies in potassium, nitrogen, and phosphorus. This paper concentrates on applying ensemble learning to predict and analyse outcomes. This paper focused on applying machine learning techniques to analyse and predict outcomes using different models, including Linear Regression for continuous predictions. Random Forest for robust classification. XGBoost for high-accuracy predictions. K-Nearest Neighbours (KNN) for pattern recognition. By testing multiple models and comparing their performance, we identified the most successful algorithm for our dataset
Optimizing Digital Marketing through Machine Learning in Cloud-Based Enterprise Systems: The Role of Web Technologies
No full textThe convergence of machine learning, cloud computing, and web technologies is transforming company strategies in digital marketing. As firms increasingly depend on data-driven tactics, machine learning provides robust capabilities for monitoring customer behavior, forecasting trends, and customizing marketing initiatives. When utilized in scalable cloud systems, these technologies provide real-time processing of extensive datasets, resulting in more flexible and efficient marketing campaigns. This article examines current developments in the incorporation of machine learning into cloud-based corporate systems, specifically highlighting its function in enhancing digital marketing. Principal topics encompass the utilization of predictive models, automation of customer interaction, and the deployment of web-based platforms to enhance data acquisition and campaign execution. Although these advances provide substantial potential, difficulties including data protection, algorithmic transparency, and system integration remain. The paper continues by delineating potential research avenues intended to tackle these issues and improve the efficacy and ethical application of machine learning in corporate marketing frameworks
Optimize Mobile App Testing Using Machine Learning to Improve User Experience
No full textAims: The study delves into the machine learning (ML) paradigm shift in enhancing mobile application testing processes for higher accuracy, efficiency, and overall user experience, with a particular focus on Decision Tree and Random Forest models.
Study Design: Experimental Research Design.
Methodology: The research applies an experimental A/B testing framework using real-world datasets and cloud-based testing environments (e.g., Firebase Test Lab) to compare ML-driven and traditional testing approaches. Techniques include automated UI defect detection through convolutional neural networks, reinforcement learning for intelligent test case prioritization, natural language processing for extracting UX-related insights from user feedback, as well as a structured user survey involving 20 participants to evaluate perceived improvements in usability and stability. MATLAB R2024b was used for model development and evaluation.
Results: Experimental results demonstrate that ML-based testing significantly outperforms traditional approaches, achieving 15–20% higher defect detection rates, 30–35% greater test coverage, and 40–50% faster execution times, alongside a notable reduction in false positives. Decision Tree and Random Forest models showed superior performance in identifying usability and performance issues. Additionally, the integration of ML into CI/CD pipelines facilitated faster bug resolution with minimal manual intervention. User survey results further confirmed improvements in user experience, with over 70% of respondents reporting enhanced application stability and responsiveness.
Conclusion: Despite its promise, deploying ML in real-world testing presents challenges, including dataset bias, variability across device environments, and limited interpretability of some model decisions. To address these, the study recommends developing robust ML-based testing frameworks, ensuring access to representative and high-quality training data, and designing hybrid models that integrate supervised learning with unsupervised anomaly detection techniques
Terror of Target Points and Loss Limits Modeling in the New York Trading Market Based on Deep Learning
No full textIn recent decades, the prediction of financial markets based on artificial intelligence has expanded a lot, which has led to the emergence of non-parametric models in this field. Models based on historical data provide traders with high accuracy predictions and do not require simplifying assumptions such as the absence of arbitrage in the market. Machine learning and especially deep learning are one of the newest topics in this field, which has been used the most in studies in recent years. According to the risk in the capital market, the use of derivative instruments, especially the option contract, is necessary for investment risk management. Forming a portfolio with the lowest available risk and with a return close to the return of the entire market is something that many investment companies are looking for. As a result, the need for a tool to predict the price of these contracts is felt, and the most important variable for pricing option contracts is implied volatility. This research is looking for a model to predict the implied volatility of option contracts using deep learning techniques, so that the prediction of this model can be used to estimate the price of option contracts. For this purpose, a modeling of target points and loss limits in the New York trading market is considered. In the proposed model, first, a probabilistic neural network with spokes, clustering operation and then classification at the data level are performed, and then the time series method based on the Brownian curve, based on control theory, can reduce dimensions, select and extract features. Based on the proposed approach, it has been shown that based on the Brownian curve, it has the ability to optimize the results of the probabilistic neural network, and the results of the combined approach, in addition to having the problem of high computational complexity, have more optimal results in terms of evaluation criteria, including accuracy
OLOREAL: A Multidisciplinary Framework for Data-driven Real Estate Innovation
No full textOLOREAL is a next-generation real estate management platform developed on the OLONIX AI platform, aimed at transforming and streamlining the property ecosystem. By incorporating cutting-edge technologies like Artificial Intelligence (AI), Machine Learning (ML), Big Data Analytics, Cloud Computing, and Blockchain, OLOREAL hopes to enhance transparency, efficiency, and decision-making for all stakeholders in the real estate sector like landlords, tenants, agents, investors, and financial institutions. The platform consists of four primary modules: PropSearch, RentLeaseShare, RealAgentCRM, and FinTechReal, each targeting critical elements in the real estate lifecycle such as property search, leasing, customer relationship management, and safe financial transactions.
With Microsoft Azure\u27s cloud infrastructure and features such as Data Lake Storage, Databricks, and Power BI, OLOREAL provides unified data ingestion, processing, and real-time visualization. Its AI-driven, blockchain-powered smart contracts deliver safe, tamper-free transactions, with AI-powered intelligent suggestions and paperless automations elevating the user experience. Through facilitation of digital engagement and contactless interaction, OLOREAL keeps pace with shifting post-pandemic business practices. With a single, data-based solution, stakeholders can streamline operations, decrease operational costs, and make more informed decisions, placing OLOREAL in a next-gen, scalable real estate platform
AI Powered Chatbot for College Information and Student Support
No full textThis study presents the development of an AI-powered chatbot tailored to assist college students with academic, campus-related, and personal development queries. Designed as a student-support software prototype, the system integrates natural language processing (NLP), text-to-speech (TTS), speech-to-text (STT), and machine learning models to deliver real-time information and interactive responses. Key features include course recommendations, timetable access, placement updates, NPTEL/hackathon alerts, library book availability, and image-based building identification. The chatbot also supports media uploads and provides mental health resources and career guidance through a conversational interface. Tested with over 50 student queries, it achieved a 92% response accuracy, identified college buildings from images with 85% accuracy, and reduced average information-seeking time by 65% compared to manual methods. Additionally, the inclusion of TTS and STT functionalities enhanced accessibility for differently abled users. This AI-driven solution fosters a smarter and more connected campus environment by streamlining support services and promoting student engagement. Future enhancements will focus on expanding the knowledge base, improving multilingual support, and refining interactive capabilities through broader user testing
Explainable AI for Livestock Disease Detection: An Integrated ML/DL Framework
No full textLivestock diseases lead to significant economic loss and threaten food security. With the increasing demand for dairy and meat products, maintaining animal health has become a critical global priority. Although farmers and agricultural workers often lack deep technical understanding of data processing, modern AI and ML technologies are now central to early disease detection in livestock. Interpretable Machine Learning (IML) and Explainable AI (XAI) provide opportunities to build trust by making model predictions transparent and understandable. This article explores XAI and IML approaches for health monitoring in farm animals, offering insights into early symptom recognition through sensor data and image analysis. XAI integrates CNN-based visual diagnostics and real-time sensor stream interpretation, while IML utilizes SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) for symptom pattern explanation and decision support. Experimental results using publicly available datasets of livestock behavior and visual symptom records demonstrate that XAI/IML-based systems provide farmers and veterinarians with clear, actionable insights to enhance livestock welfare and productivity
Secure Front-End Automation Framework: A Novel Approach to Client-Side Data Encryption and Zero Trust API Interaction
No full textThis study aims to evaluate the effectiveness of the Secure Front-End Automation Framework (SFAF) in enhancing front-end application security and performance compared to traditional web development frameworks. The focus is on client-side encryption and Zero Trust API interactions .and study design was Experimental research design. Regarding to study place Smart barrel, Miami, United States, from September 2024 to March 2025.In this study methodology two web applications were developed. One used a conventional client-server model with standard security protocols, while the other implemented SFAF with advanced client-side encryption and Zero Trust-based API interactions. Automated security testing tools such as OWASP ZAP, Burp Suite, and Postman were used to collect data from 60 test instances (30 per group). Key performance indicators included response time, memory usage, CPU load, unauthorized API call attempts, and compliance with OWASP Top 10 security benchmarks. Statistical analysis was conducted using paired-samples t-tests, independent-samples t-tests, and Cohen’s d for effect size. Controlled simulations ensured high internal validity. Commonly exploited web scenarios were used to enhance external validity Applications based on SFAF showed a statistically significant reduction in unauthorized API interaction attempts (p < 0.01) and a 35% improvement in compliance with OWASP Top 10 benchmarks compared to traditional applications. Although a slight increase in average response time (2.7%) and resource consumption was observed, these differences were statistically insignificant (p > 0.05). Effect size calculations further confirmed the practical significance of the results. The proposed Secure Front-End Automation Framework combines client-side data encryption with Zero Trust API interactions, offering robust security measures that aid industries in meeting critical regulatory compliance standards such thereby enhancing data privacy and minimizing risks associated with unauthorized access. The Secure Front-End Automation Framework significantly enhances front-end security without substantially affecting system performance. It offers a viable solution for developing scalable, Zero Trust-compliant web applications. These findings support adopting SFAF as a foundational approach to modern web application security in response to emerging threats. This study contributes to the academic understanding of client-side security by integrating decentralized encryption models with Zero Trust architecture for developers and policymakers
Semantic Search for Data on a Given Topic in Social Networks: A Comparative Study of Keyword-based and BERT-based Methods
No full textSemantic search has emerged as a powerful alternative to traditional keyword-based retrieval, particularly in the context of unstructured social media data. This study presents a comparative analysis of a semantic search system based on Sentence-BERT (SBERT) and a conventional keyword-based pipeline implemented with Elasticsearch, using a large Reddit dataset as a case study. The primary contribution lies in integrating state-of-the-art semantic modeling with scalable search infrastructure and empirically evaluating its effectiveness on real-world social media content. The experimental workflow includes six stages: dataset selection, preprocessing, embedding generation, indexing, query processing, and performance evaluation. Results show that the SBERT-based semantic search system consistently outperforms the keyword-based approach across all metrics, particularly in capturing user intent, handling informal language, and retrieving semantically relevant content despite lexical variations. Nonetheless, the semantic approach incurs higher computational costs and exhibits occasional overgeneralization