IAES International Journal of Artificial Intelligence (IJ-AI)
Not a member yet
1769 research outputs found
Sort by
Automated menu planning for pregnancy based on nutrition and budget using population-based optimization method
Get PDFNutritional fulfilment during pregnancy depends on the budget. Meanwhile, nutrition is needed during pregnancy to keep the mother and fetus healthy. Therefore, this study aims to assist maternal nutrition planning by using population-based optimization methods such as genetic algorithm (GA), particle swarm optimization (PSO), duck swarm algorithm (DSA), and whale optimization (WO) according to their nutritional needs at minimum cost. Additionally, this study compares the method performance to find the best method. There are 55 foods obtained from previous studies divided into five groups: staple food (SF), vegetables (VG), plant-source food (PS), animal-source food (AS), and complementary (CP). The model evaluation results show that GA's performance differed significantly from other models because it obtained the highest fitness by 439.73 and more variation in fitness results. Three models other than GA have no significant difference, but DSA performance obtained a superior fitness of 367.18. Furthermore, optimization methods must be combined with other artificial intelligence methods to develop innovative technology to support maternal nutrition and prevent stunting
Artificial intelligence of things: society readiness
Get PDFThe convergence of artificial intelligence (AI) and the internet of things (IoT), known as the artificial intelligence of things (AIoT), represents a transformative leap in technology. This study investigated societal readiness for AIoT adoption and identified key factors influencing the readiness. The researchers used technology readiness index (TRI) model and broken down the model into the online survey’s instrument. The study used about 129 samples for examining the used variables, i.e., perceptions of innovation, technological skills, social and cultural influences, regulatory factors, and digital literacy. The authors employed partial least squares structural equation modeling (PLS-SEM) method using SmartPLS 3.0 to analyze the relationships between the variables of the model. The results highlighted innovation as a significant driver of societal readiness, while factors like discomfort have a lesser impact. Security and optimism also played moderate roles in shaping readiness. These findings offer crucial insights for stakeholders of the AIoT implementation by providing a foundation for strategies that promote the successful integration of AIoT into society. The study contributes to the broader discourse on technology adoption, offering a roadmap for enhancing societal preparedness
Optimized ensemble modeling approach for student cumulative grade point average prediction using regression models
Get PDFThis research focuses on developing models to accurately predict student’s cumulative grade point average (CGPA) in the early stages of their study to tackle the problem of dropout rates in educational institutions. The state-of-the-art methods address CGPA prediction as a classification problem, providing only an approximate prediction where precise prediction is essential. In this research, six regression models, namely linear regression, support vector regression (SVR), decision tree (DT), random forest (RF), lasso regression (LR), and ridge regression (RR) are developed without optimization and later fine-tuned using Bayesian optimization (BO) and GridSearchCV. BO efficiently searches the hyper-parameter space using probabilistic distribution’s function, whereas GridSearchCV exhaustively searches the hyper-parameter space. These techniques significantly improved the model's performance; SVR achieved an R² score of 94.11% through BO. Ensemble techniques, such as stacking, voting, and boosting, can further enhance the predictive capability of the model. The stacking ensemble model achieved the highest R² score of 94.45%, providing a 0.50% improvement in the R2 score. The findings of this study suggest that advanced optimization and ensemble techniques can substantially enhance the predictive capability of the model, thus enabling institutions to support students at risk of academic probation proactively
BonoNet: a deep convolutional neural network for recognizing bangla compound characters
Get PDFThe bangla alphabet includes vowels, consonants, and compound symbols. The compound nature of bangla is a product of combining two or more root bangla characters into one graph. They are difficult to differentiate because they have a sophisticated geometric shape and an immense variety of scripts used by different places and individuals. This is one of the greatest challenges in creating effective optical character recognition (OCR) systems for bangla. In this paper, a deep convolutional neural network (DCNN)-based system is presented to identify bangla compound characters with high precision. The model was trained using the AIBangla dataset. It has about 171 classes of bangla compound characters. A DCNN system, BonoNet, was designed to classify compound characters. BonoNet outperformed all other state-of-the-art architecture on the test set and improved over current state-of-the-art architecture methods. BonoNet will greatly improve the automation and analysis of the bangla language by accurately identifying these compound complex characters
Facial features extraction using active shape model and constrained local model: a comprehensive analysis study
Get PDFHuman facial feature extraction plays a critical role in various applications, including biorobotics, polygraph testing, and driver fatigue monitoring. However, many existing algorithms rely on end-to-end models that construct complex classifiers directly from face images, leading to poor interpretability. Additionally, these models often fail to capture dynamic information effectively due to insufficient consideration of respondents' personal characteristics. To address these limitations, this paper evaluates two prominent approaches: the constrained local model (CLM), which accurately extracts facial features depending on patch experts, and the active shape model (ASM), designed to simultaneously extract the appearance and shape of an object. We assess the performance of these models on the MORPH dataset using point to point error as evaluation metrics. Our experimental results demonstrate that the CLM achieves higher accuracy, while the ASM exhibits better efficiency. These findings provide valuable insights for selecting the appropriate model based on specific application requirements
An energy-efficient and secure framework for wireless sensor networks
Get PDFIn wireless sensor networks (WSNs), achieving energy efficiency, security, and minimizing route change propagation time is essential for maintaining optimal performance. This paper introduces a new approach that combines Bray Jaccard Curtis-based Calinski Harabasz k-means (BJC-CHKMeans) for clustering and Karl Pearson correlation-based egret swarm optimization algorithm (KPC-ESOA) for selecting the best cluster head (CH) and path, along with classifying long short-term memory with gated recurrent units (CLE-GRU) for detecting harmful nodes. The methodology aims to enhance energy usage, improve routing efficiency, and strengthen security by identifying malicious nodes. Additionally, it integrates a secure routing table using elbow de-swinging k-anonymity (EDS-KA) and employs digital signature algorithm-based Zeta Bernoulli Merkle tree (DSA-ZBMT) to ensure secure communication with sink nodes. The WSN-DS dataset was used for training and testing, with rigorous preprocessing, feature extraction, and selection to maintain data integrity. Experimental results revealed that the proposed BJC-CHKMeans and CLE-GRU models outperform traditional methods in power consumption, latency, and accuracy. The system achieved a power consumption of 2.1 mW for clustering and 1.9 mW for classification, while also providing near-perfect accuracy in detecting harmful nodes. These findings demonstrate that the framework significantly enhances the energy efficiency and security of WSNs, making it a highly effective solution for large, dynamic sensor networks
Modeling sentiment analysis of Indonesian biodiversity policy Tweets using IndoBERTweet
Get PDFThis study develops and evaluates a sentiment analysis model using IndoBERTweet to analyze Twitter data on Indonesia’s biodiversity policy. Twitter data focusing on topics such as food security, health, and environmental management were collected, with a representative subset of 13,435 tweets annotated from a larger dataset of 500,000 to ensure reliable sentiment labels through majority voting. IndoBERTweet was compared to seven traditional machine-learning classifiers using TF-IDF and BERT embeddings for feature extraction. Model performance was assessed using mean accuracy, mean F1 score, and statistical significance (p-values). Additionally, sentiment analysis included word attribution techniques with BERT embeddings, enhancing relevance, interpretability, and consistent attribution to deliver accurate insights. IndoBERTweet models consistently outperformed traditional methods in both accuracy and F1 score. While BERT embeddings boosted performance for conventional models, IndoBERTweet delivered superior results, with p-values below 0.05 confirming statistical significance. This approach demonstrates that the model’s outputs are explainable and align with human understanding. Findings underscore IndoBERTweet’s substantial impact on advancing sentiment analysis technology, showcasing its potential to drive innovation and elevate practices in the field
A comprehensive artificial intelligence framework for reducing patient rehospitalizations
Get PDFThe role of artificial intelligence (AI) in the healthcare sector is increasing daily. Readmissions of patients have become a significant challenge for the medical sector, adding unnecessary burden. Governments and public sectors are continuously working on the hospital readmissions reduction program (HRRP). In this research work, an AI framework has been developed to reduce patient readmissions. The accuracy of the framework has been increased by continuous refinement in feature engineering, incorporating several complex datasets. The framework analyses the different algorithms like bidirectional long short-term memory (BiLSTM), convolutional neural network (CNN), and XGBoost for prediction. This framework has shown a 92% accuracy rate during testing, showing a 37% reduction in 40-day rehospitalization rates. This reduces the overburden on hospital systems by avoiding unnecessary readmissions of patients. The system’s real-time development, scalability, management of things in an ethical manner, and long-term viability will remain as future scope
Anisa: artificial intelligence companion for elderly care with empathetic conversations and health management
Get PDFThis study introduces Anisa, an advanced artificial intelligence (AI) companion designed to enhance elderly care by addressing the multifaceted needs and challenges of older adults. The system integrates the Llama 3.2 model, powered by Groq, to facilitate context-aware dialogues and empathetic interactions. This capability helps alleviate loneliness and provides essential companionship. Agenda.js is used for scheduling and managing reminders, ensuring timely notifications for medications and appointments. Additionally, Twilio enables emergency alerts when distress signals are detected. Anisa promotes physical activity, tracks daily routines, and generates activity reports shared with caregivers and healthcare providers. Expo CLI implements step-tracking and document-sharing features. By integrating these functionalities, Anisa improves the quality of life for seniors, eases caregiver responsibilities, and fosters a safer, more supportive environment
Accelerating solder joint classification using generative artificial intelligence for data augmentation
Get PDFDespite advancements in computer vision, deploying deep learning algorithms for automated optical inspection (AOI) in printed circuit board (PCB) manufacturing remains challenging due to the need for large, diverse, and high-quality training datasets. AOI programs must be developed quickly, often as soon as the first PCB is assembled, to meet tight production timelines. However, deep learning models require extensive datasets of defect images, which are both scarce and time-consuming to collect. As a result, AOI software developers frequently resort to traditional rule-based methods. This study introduces a novel framework that leverages generative AI and discriminative AI to address dataset limitations. By applying a diffusion model to systematically add and remove Gaussian noise, the framework generates realistic defect images, expanding the available training data. This data augmentation accelerates the learning process of deep learning models, enhancing their robustness and generalizability. Experimental results demonstrate that this approach improves AOI system performance by producing balanced datasets across various defect classes. The framework shortens training times while maintaining high inspection accuracy, facilitating faster deployment of AOI systems in manufacturing. This advancement enhances quality control processes, contributing to more efficient, and reliable mass production of PCBs