International Journal of Advances in Artificial Intelligence and Machine Learning
The International Journal of Advances in Artificial Intelligence and Machine Learning (IJAAIML) is a prominent academic journal dedicated to publishing cutting-edge research and developments in the fields of Artificial Intelligence (AI) and Machine Learning (ML). It serves as an essential platform for researchers, practitioners, and professionals worldwide to share innovative ideas, technologies, and empirical studies that contribute to advancing AI and ML. The journal emphasizes both theoretical advancements and practical applications, showcasing how these technologies are shaping various industries, including healthcare, finance, education, robotics, and autonomous systems. IJAAIML covers a wide range of topics within AI and ML, such as deep learning, neural networks, natural language processing (NLP), computer vision, robotics, data mining, reinforcement learning, and AI ethics. The journal is open to diverse types of scholarly contributions, including original research articles, review papers, case studies, technical notes, and surveys. It encourages submissions that introduce novel algorithms, methodologies, and systems, as well as those addressing challenges and proposing new approaches in AI and ML. This broad scope allows the journal to remain at the forefront of technological innovation, providing valuable insights into the latest trends and developments in the field. The journal maintains high academic standards through a rigorous peer-review process, ensuring that each published article is of exceptional quality and originality. Submissions are evaluated by experts in relevant fields based on their significance, innovation, methodology, and clarity. This commitment to quality makes IJAAIML a trusted source of information for a diverse audience, including academic researchers, industry professionals, AI practitioners, and students who seek to stay informed about the latest advances in AI and ML. IJAAIML is committed to global knowledge dissemination, making its publications accessible to researchers and professionals worldwide through its online platform. This approach fosters knowledge exchange and collaboration across borders, enabling the journal to reach a broad international audience. By highlighting state-of-the-art research that addresses real-world problems using AI and ML technologies, IJAAIML plays a significant role in advancing the understanding and application of these technologies. Additionally, the journal explores the ethical, societal, and economic impacts of AI and ML, promoting discussions on responsible AI practices and future directions. By contributing to these conversations, IJAAIML not only advances technological innovation but also encourages the development of AI and ML in a manner that considers broader implications for society. Overall, the International Journal of Advances in Artificial Intelligence and Machine Learning stands as a crucial resource for anyone involved in AI and ML, driving forward the frontiers of these dynamic fields through high-quality, peer-reviewed research.
Articles
27 Documents
<![CDATA[A Deep Learning Approach to Sentiment Analysis of Hotel Reviews: Comparing BERT and LSTM Models ]]>
<![CDATA[Wang, Gunawan]]>;
<![CDATA[Jaber, Mustafa Musa]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.403
<![CDATA[Background of study: Online reviews are crucial in shaping consumer decisions, particularly in the hospitality industry, but accurately extracting sentiment remains challenging due to language subjectivity, varied expression styles, and significant class imbalance where positive reviews outweigh negative and neutral ones. Traditional machine learning methods often fail to address these issues effectively, favoring the majority class.Aims and scope of paper: This study employs BERT and LSTM deep learning models to classify hotel reviews into positive, neutral, and negative sentiment categories. The primary aim is to compare their performance in sentiment analysis and managing imbalanced data, evaluating both with and without under-sampling.Methods: A 20,000-review dataset from TripAdvisor was pre-processed, including stop word/special character removal, tokenization, stemming, and lemmatization. Star ratings were categorized: 4-5 as positive, 3 as neutral, and 1-2 as negative. Random under-sampling was applied to the majority (positive) class to balance the dataset. BERT (bert-base-uncased) and LSTM models were trained with an 80:20 training-validation split, and evaluated using accuracy, precision, recall, and F1-score, with 5-fold cross-validation.Result: BERT without under-sampling achieved the highest overall accuracy of 0.86, with strong F1-scores for positive (0.93) and negative (0.79) sentiments. However, all models struggled with neutral sentiments (BERT F1-score: 0.43, LSTM: 0.25). Under-sampling improved neutral class recall (BERT: 0.79) but decreased overall accuracy (BERT: 0.73; LSTM: 0.67) and positive class precision.Conclusion: BERT generally outperforms LSTM for hotel review sentiment analysis, particularly with imbalanced data. While under-sampling helps address class imbalance by improving neutral recall, it incurs significant performance trade-offs, reducing overall accuracy and precision in majority classes due to information loss. Future work should explore advanced resampling (SMOTE, ADASYN) or transfer learning (RoBERTa, XLNet) for better balance and neutral sentiment classification.]]>
<![CDATA[Comparative Study of CNN and Vision Transformers on Indonesian Tradisional Cakes Classification ]]>
<![CDATA[Trisnawarman, Dedi]]>;
<![CDATA[Supriyanton, Adolf Asih]]>;
<![CDATA[Mawardi, Viny Christanti]]>;
<![CDATA[Okengwu, Ugochi A]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.405
<![CDATA[Background of study: Food image classification is a challenging task in computer vision, particularly when dealing with traditional food items that exhibit subtle visual variations. While Convolutional Neural Networks (CNNs) have long been the standard for image recognition, their limitation in capturing long-range dependencies has led to the emergence of Vision Transformers (ViTs). In this context, the classification of Indonesian traditional cakes offers a culturally rich yet complex problem for automated image recognition systems.Aims and scope of paper: This study aims to conduct a comparative analysis between EfficientNet-B0 (CNN-based) and ViT-B/16 (Transformer-based) architectures in classifying eight categories of Indonesian traditional cakes. The research evaluates not only classification accuracy but also the strengths and limitations of each model in handling fine-grained visual distinctions.Methods: Both models were fine-tuned using the “Kue Indonesia” dataset from Kaggle. The methodology includes image preprocessing, model training with consistent parameters, and evaluation using accuracy, precision, recall, and F1-score. A confusion matrix was also used to visualize misclassifications and analyze per-class performance.Result: ViT-B/16 achieved slightly higher accuracy (96.25%) compared to EfficientNet-B0 (95.62%). ViT performed better in classes with subtle variations, such as kue lapis and kue dadar gulung, while EfficientNet-B0 showed superior efficiency and high accuracy on visually distinct cakes.Conclusion: Both CNN and ViT models demonstrate strong performance in traditional food classification. ViT is more robust in fine-grained visual analysis, whereas EfficientNet-B0 is preferable for resource-constrained environments. This study highlights the role of AI in supporting digital preservation of culinary heritage.]]>
<![CDATA[3D Box Packing with Heuristics and Metric Analytics ]]>
<![CDATA[Kasem Alqudah, Mashal]]>;
<![CDATA[Pambudi, Dhidhi]]>;
<![CDATA[Zakaria, Mohd Zaki]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.409
<![CDATA[Background of Study: The 3D Bin Packing Problem (3D-BPP) is an NP-hard problem crucial for logistics and supply chain optimization, aiming to efficiently pack boxes into containers while maximizing space and maintaining stability. Traditional heuristics like First Fit and Best Fit are fast but lack optimality and adaptability in dynamic environments. Metaheuristic approaches, such as Genetic Algorithms (GA), offer better solutions but with higher computational costs.Aims and Scope of Paper: This study presents a comparative analysis of First Fit, Best Fit, and a custom Genetic Algorithm as packing strategies for 3D-BPP. It evaluates these methods against multiple performance metrics to understand their trade-offs and proposes future research directions.Methods: The study uses a dataset of 5,000 cargo records from an Indonesian logistics company, including item dimensions and weights, preprocessed for normalization and filtering. A 3D simulation environment built with PyBullet visualizes the packing process. Performance metrics include space utilization, total packed weight, packing time, access efficiency, stability score, and placement success rate. A Wall-Building heuristic acts as a fallback for unplaced items.Result: First Fit provides fast, lightweight solutions suitable for real-time applications. Best Fit shows marginally better space utilization but lacks robustness. The Genetic Algorithm outperforms both heuristics in packing quality, accessibility, and load stability, though with significantly higher computation time. No single algorithm dominates across all metrics.Conclusion: The choice of packing method should align with specific operational constraints: speed, compactness, or quality. A hybrid model combining heuristic initialization with GA refinement is a promising direction for future research to develop more intelligent, context-aware packing systems.]]>
<![CDATA[Analyzing Bias in Large Language Models: A Quantitative Study Using Sentiment and Demographic Metrics ]]>
<![CDATA[Mandava, Ramya]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.411
<![CDATA[Background of study: The widespread adoption of Large Language Models (LLMs) raises concerns about biases that affect fairness and credibility. As LLMs affect areas such as recruitment and customer service, systematic quantitative analysis is essential to identify and mitigate these biases.Aims and scope of paper: This research investigates demographic bias in LLM quantitatively by analyzing sentiment polarity scores across different demographic categories. The goal is to provide a statistically confirmed analysis of sentiment bias and propose mitigation methods, focusing on GPT-4, LLaMA-2, Claude, and BLOOM.Methods: Quantitative analysis was performed on GPT-4, LLaMA-2, Claude, and BLOOM using sentiment and demographic data. Sentiment polarity assessments for gender and racial/ethnic groups were obtained with VADER and TextBlob. Demographic Disparity Score, ANOVA, and Cohen's Kappa assessed the significance and appropriateness of bias. Inter-rater reliability between automated tools and human annotators was also evaluated.Result: Sentiment bias was found in all models, varying by gender and race, particularly in GPT-4 and Claude. Sentiment scores were consistently higher for queries pertaining to females than those pertaining to males across all models, with GPT-4 and Claude showing the largest differences. Claude also showed racial sentiment alignment, favoring queries relating to white people over black people. ANOVA confirmed statistically significant sentiment variation by demographics across all models. High inter-rater reliability validated the sentiment analysis.Conclusion: This study shows demographic bias in GPT-4, LLaMA-2, Claude, and BLOOM, with different sentiment trends across demographic classifications. The models showed more positive sentiment for female questions and a trend towards certain racial groups. These findings indicate an embedded bias in the training data, which raises ethical concerns. Identifying and addressing these biases is critical to ensuring fairness and credibility in real-world LLM applications.]]>
<![CDATA[Hepatitis Disease Prediction Using Convolutional Neural Network Algorithm in Machine Learning Technology ]]>
<![CDATA[Sirisati, Ranga Swamy]]>;
<![CDATA[Jayasri, B.]]>;
<![CDATA[Avanthi, A.]]>;
<![CDATA[Ramyasri, A.]]>;
<![CDATA[Sowmya, K.]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.419
<![CDATA[Background of Study: Hepatitis is a significant viral infection causing liver inflammation, potentially leading to hepatocyte death and impaired liver function. Types B (HBV) and C (HCV) can cause chronic hepatitis, cirrhosis, and cancer. Globally, around 257 million people are infected with HBV and 71 million with HCV. Early detection of chronic Hepatitis B is crucial for effective management.Aims and Scope of Paper: This study aims to predict hepatitis progression in patients from their medical histories. It seeks to enhance prediction accuracy by addressing challenges like noise and inefficiency caused by similar aspect values and distributions within datasets.Methods: Machine learning, a branch of AI, is employed for chronic disease prediction. The study primarily utilizes the K-Nearest Neighbour (KNN) algorithm to predict and eliminate redundant data and noise. Other models evaluated include Logistic Regression, Random Forest, and Convolutional Neural Networks (CNN), with SMOTE used for dataset balancing.Result: KNN achieved 0.970 accuracy, Logistic Regression 0.966, and Random Forest 0.95. The CNN model demonstrated exceptional performance, reaching 1.0 accuracy with perfect precision, recall, and F1-score for Hepatitis A and B.Conclusion: While KNN performed well among traditional methods, deep learning models like CNN show superior accuracy and generalizability, offering a robust framework for hepatitis prediction.]]>
<![CDATA[Performance Evaluation of YOLOv10 and YOLOv11 on Blood Cell Object Detection Dataset ]]>
<![CDATA[Džakula, Nebojša Bačanin]]>;
<![CDATA[Heriansyah, Rudi]]>;
<![CDATA[Fadly, Fadly]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.434
<![CDATA[Background of study: Blood cell analysis is vital for diagnosing medical conditions, but traditional manual methods are laborious and error-prone. Deep learning, especially YOLO models, offers automated solutions for medical image analysis. However, the real-world effectiveness of the latest YOLOv11 in blood cell detection is not thoroughly investigated, as general object detection improvements may not translate to biomedical images due to their unique characteristics.Aims and scope of paper: This study systematically compares YOLOv10 and YOLOv11 on a public blood cell detection dataset to assess if YOLOv11's advancements provide tangible benefits for blood cell classification. The goal is to identify the most effective model for accurate and efficient detection in microscopic images, guiding AI-driven diagnostic tool selection.Methods: Both models were trained and tested under identical conditions using the Kaggle Blood Cell Detection Dataset (RBCs, WBCs, Platelets). Images were resized to 640x640 pixels. Performance metrics included mAP (mAP@50 and mAP@50–95), Precision, Recall, F1-score, speed, model complexity, and training time.Result: YOLOv11n consistently showed higher accuracy (mAP50: 0.9279 vs. 0.9120; mAP50-95: 0.6524 vs. 0.6347), particularly for RBCs and WBCs. However, YOLOv11n had longer inference (11.35 ms/image) and postprocessing times (8.64 ms/image) compared to YOLOv10n (7.00 ms/image and 0.90 ms/image). YOLOv11n trained faster (0.311 hours vs. 0.375 hours), with a smaller model size (5.5 MB vs. 5.8 MB), fewer parameters, and reduced computational complexity.Conclusion: YOLOv11n offers superior accuracy and improved training efficiency, making it suitable for medical image object detection where precision is paramount. The increased inference and postprocessing times indicate a performance-speed trade-off. Model selection should balance these factors based on deployment context.]]>
<![CDATA[A Comparative Study of Convolutional Neural Networks and Vision Transformers for Fruit Classification ]]>
<![CDATA[Jawarneh, Malik]]>;
<![CDATA[Marwanto, Arief]]>;
<![CDATA[Syamsuar, Dedy]]>;
<![CDATA[Kusnandar, Maivi]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.435
<![CDATA[Background of study: Accurate fruit classification is vital for agricultural automation, yet traditional methods are often subjective and inefficient. Convolutional Neural Networks (CNNs) are effective but struggle with global context in fine-grained tasks. Vision Transformers (ViTs), inspired by NLP models, offer global attention mechanisms that may improve classification in complex scenarios.Aims and scope of paper: This study compares the performance of EfficientNet-B0 (a CNN model) and ViT-B/16 (a Transformer model) on a fruit classification task involving five fruit types. The goal is to evaluate their strengths and weaknesses under controlled experimental conditions using a moderately sized dataset.Methods: A dataset of 10,000 fruit images was preprocessed with standard augmentation techniques and split into training and validation sets. Both models were fine-tuned using pretrained weights. Performance was evaluated using accuracy, precision, recall, F1-score, and confusion matrices.Result: EfficientNet-B0 achieved higher overall accuracy (94%) than ViT-B/16 (92%). The CNN model performed consistently across all classes, particularly excelling in bananas and strawberries. ViT-B/16 showed superior results for strawberries but struggled with apples. Confusion matrices revealed class-specific strengths and weaknesses.Conclusion: EfficientNet-B0 is better suited for general fruit classification due to its balanced performance, while ViT-B/16 excels in capturing fine-grained visual features. A hybrid approach may leverage both models’ strengths for enhanced performance in real-world applications.]]>
<![CDATA[Development of a WebXR-Based Collaborative LMS System with 3D Virtual Features and Artificial Intelligence ]]>
<![CDATA[Padmasari, Ayung Candra]]>;
<![CDATA[Azizan, Ahamad Tarmizi]]>;
<![CDATA[Ergashevna, Burieva Kibrio]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.448
<![CDATA[Background of Study: The advancement of immersive technology and artificial intelligence (AI) offers new opportunities for creating more adaptive and interactive learning systems. However, higher education institutions still face challenges such as limited industry-standard facilities and the high cost of multimedia equipment.Aims and Scope of Paper: This study aims to develop a prototype of a WebXR-based Collaborative Learning Management System (LMS) equipped with 3D virtual features and AI integration to enhance student learning experiences.Methods: The research employed the Multimedia Development Life Cycle (MDLC) method, which consists of six stages: conceptualization, design, material collection, assembly, testing, and distribution. The study involved 30 Multimedia Education students from Universitas Pendidikan Indonesia selected through purposive sampling.Result: Feasibility testing using a Likert-scale questionnaire revealed that the system achieved a highly feasible category with average scores of 81,7 % for Learnability 85,6 %, for system performance 76,93%, for Efficiency 79,9%, for memorability 74,2%, satisfaction 85,4 %. Resulting in an overall feasibility of 81.7%. Semi-structured interviews confirmed that AI integration significantly supported learning personalization and provided content recommendations, although the AI feature was limited to text-based responses.Conclusion: The results indicate that combining WebXR and AI in an LMS can address the challenges of industry-based learning by providing immersive, adaptive, and accessible learning experiences. This system demonstrates strong potential as a future-ready digital learning solution, with future research suggested to evaluate its impact on learning outcomes and improve AI capabilities for deeper contextual interaction.]]>
<![CDATA[Rice Grain Quality Analysis Using Image Processing ]]>
<![CDATA[Rani, K.Sandhya]]>;
<![CDATA[Swetha, K]]>;
<![CDATA[Varshini , K. Amrutha]]>;
<![CDATA[Harika, G]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.455
<![CDATA[Background of study: Rice quality is crucial for global food security and market value, but traditional assessment relies on labor-intensive, inconsistent, and error-prone manual inspection.Aims and scope of paper: This research proposes an automated system using image processing and AI for comprehensive rice grain quality analysis. The goal is to develop a robust, objective, and precise system to classify rice varieties and evaluate quality with minimal human intervention, reducing the effort, cost, and time of traditional methods.Methods: The core contribution is a computerized model that uses digital image processing to automatically segment, identify, and extract key quality parameters like length, width, area, perimeter, and shape descriptors. The methodology includes image acquisition, preprocessing (binary conversion, thresholding, noise reduction, morphological operations), edge detection, feature extraction (especially aspect ratio), classification, and visualization. The system was trained on a self-curated dataset of various rice varieties.Result: The system successfully analyzed Sona Masuri, Basmati, and Jasmine rice varieties based on grain count and average aspect ratio. Sona Masuri (211 grains, 1.57 aspect ratio) and Basmati (261 grains, 1.8 aspect ratio) were classified as 'Bold'. Jasmine (30 grains, 2.1 aspect ratio) was classified as 'Medium' , consistent with defined criteria.Conclusion: The project successfully analyzed and processed rice grain images to determine size, shape, and quality, accurately measuring length, width, and aspect ratio for classification. Image processing techniques improved image quality and defect detection. The system objectively applied classification logic, demonstrating high precision in rice grain quality determination, which is crucial for market value and consumer satisfaction.]]>
<![CDATA[Innovative Blood Group Detection Through Image Processing and FingerPrint Recognition ]]>
<![CDATA[Eenaja, Aileni]]>;
<![CDATA[Gunda, Rishitha]]>;
<![CDATA[Ashwini, Kasineedi]]>;
<![CDATA[Keerthi, P]]>;
<![CDATA[Sravani, Naika]]>
<![CDATA[ International Journal of Advances in Artificial Intelligence and Machine Learning Vol. 2 No. 2 (2025): International Journal of Advances in Artificial Intelligence and Machine Learni ]]>
Publisher : <![CDATA[CV Media Inti Teknologi ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.58723/ijaaiml.v2i2.458
<![CDATA[Background of study: Traditional blood group determination methods are time-consuming, invasive, and require specialized equipment and trained personnel, leading to delays in medical decisions in remote or emergency settings.Aims and scope of paper: This project explores an innovative, non-invasive approach to blood group detection using fingerprint recognition and image processing, aiming to overcome limitations of prior methods regarding accuracy, scalability, and accessibility. The core hypothesis is that unique fingerprint patterns can correlate with blood groups using advanced machine learning.Methods: The proposed system involves fingerprint image acquisition (via smartphone/scanner), pre-processing (noise reduction, grayscale conversion, etc.), feature extraction using ORB and GLCM, and classification with a Convolutional Neural Network (CNN). The lightweight MobileNet architecture is utilized for efficiency, trained on a self-collected dataset of 60,000 thumb images categorized into 8 blood group classes, with HOG integrated for enhanced feature extraction. The system is accessible via a user-friendly chatbot interface.Result: Experimental evaluation demonstrated robust performance across various deep learning models. ResNet50 achieved the highest accuracy of 95.3% on the BloodHub Dataset. The custom CNN model achieved 94.8% accuracy on the Custom Fingerprint Dataset, and MobileNet achieved a commendable 93.6% accuracy on the BloodCell-Detection-Dataset.Conclusion: This project presents a viable, non-invasive blood group detection method by combining fingerprint biometrics, advanced image processing, and deep learning within a chatbot interface. Deep architectures like ResNet50 and the tailored CNN consistently achieved over 94% accuracy, validating the feasibility of reagent-free, portable blood typing for emergency, rural, and resource-constrained environments. This system can democratize critical diagnostic services and enhance patient care.]]>
