IAES International Journal of Artificial Intelligence (IJ-AI)
Not a member yet
1769 research outputs found
Sort by
Traffic flow prediction using long short-term memory-Komodo Mlipir algorithm: metaheuristic optimization to multi-target vehicle detection
Get PDFMulti-target vehicle detection in urban traffic faces challenges such as poor lighting, small object sizes, and diverse vehicle types, impacting traffic flow prediction accuracy. This study introduces an optimized long short-term memory (LSTM) model using the Komodo Mlipir algorithm (KMA) to enhance prediction accuracy. Traffic video data are processed with YOLO for vehicle classification and object counting. The LSTM model, trained to capture traffic patterns, employs parameters optimized by KMA, including learning rate, neuron count, and epochs. KMA integrates mutation and crossover strategies to enable adaptive selection in global and local searches. The model's performance was evaluated on an urban traffic dataset with uniform configurations for population size and key LSTM parameters, ensuring consistent evaluation. Results showed LSTM-KMA achieved a root mean square error (RMSE) of 14.5319, outperforming LSTM (16.6827), LSTM-improved dung beetle optimization (IDBO) (15.0946), and LSTM-particle swarm optimization (PSO) (15.0368). Its mean absolute error (MAE), at 8.7041, also surpassed LSTM (9.9903), LSTM-IDBO (9.0328), and LSTM-PSO (9.0015). LSTM-KMA effectively tackles multi-target detection challenges, improving prediction accuracy and transportation system efficiency. This reliable solution supports real-time urban traffic management, addressing the demands of dynamic urban environments
Artificial intelligence predictive modeling for educational indicators using data profiling techniques
Get PDFIn Morocco, the escalating challenges in the education sector underscore the necessity for precise predictions and informed decision-making. Effective management of the education system depends on robust statistical data, which is crucial for guiding decisions, refining policies, and improving both the quality and accessibility of education. Reliable indicators are vital for ensuring efficiency, equity, and accuracy in educational planning and decision- making. Without dependable data, implementing effective policies, addressing the needs appropriately, and achieving positive outcomes becomes difficult. This paper aims to identify the optimal machine learning model for analyzing educational indicators by comparing a range of advanced models across a comprehensive set of metrics. The objective is to determine the most effective model for profiling relevant information and addressing predictive challenges with high accuracy
Enhancing challenge-based immersion in cultural game using appreciative fuzzy logic
Get PDFMany traditional games in Indonesia are considered cultural heritage and are in serious decline; young generations no longer know about them. Serious games have been considered a potential educational tool for cultural heritage preservation. Lack of immersive experience due to over-focus on the learning content is a common problem in those games. Very little research also discusses cultural heritage serious game design frameworks. This study uses the appreciative fuzzy logic system (AFLS) to enhance the challenge-based immersive experience (CBIE) in the Joglosemar cultural heritage game. The AFLS provides autonomous challenges, such as enemy numbers and aggressive behavior, and the frequency of item appearances in the games using fuzzy logic with respect to the appreciative serious games (ASG) concepts. The ASG is the design guide for serious games that divides the game activities into 4-D: discovery, dream, design, and destiny. We use three ASG-based serious games to evaluate the CBIE produced by AFLS. The game experience questionnaire (GEQ) is used to measure the player experience, while the cross-validation is used to measure the AFLS performance. Results show that the AFLS enhances the CBIE. The study contributes mainly to provide reliable intelligent system for automated serious game design
Contract-based federated learning framework for intrusion detection system in internet of things networks
Get PDFA plethora of national vital infrastructures connected to internet of things (IoT) networks may trigger serious data security vulnerabilities. To address the issue, intrusion detection systems (IDS) were investigated where the behavior and traffic of IoT networks are monitored to determine whether malicious attacks or not occur through centralized learning on a cloud. Nonetheless, such a method requires IoT devices to transmit their local network traffic data to the cloud, thereby leading to data breaches. This paper proposes a federated learning (FL)-based IDS on IoT networks aiming at improving the intrusion detection accuracy without privacy leakage from the IoT devices. Specifically, an IoT service provider can first motivate IoT devices to participate in the FL process via a contract-based incentive mechanism according to their local data. Then, the FL process is executed to predict IoT network traffic types without sending IoT devices’ local data to the cloud. Here, each IoT device performs the learning process locally and only sends the trained model to the cloud for the model update. The proposed FL-based system achieves a higher utility (up to 44%) than that of a non-contract-based incentive mechanism and a higher prediction accuracy (up to 3%) than that of the local learning method using a real-world IoT network traffic dataset
An algorithm for training neural networks with L1 regularization
Get PDFThis paper presents a new algorithm for building neural network models that automatically selects the most important features and parameters while improving prediction accuracy. Traditional neural networks often use all available input parameters, leading to complex models that are slow to train and prone to overfitting. The proposed algorithm addresses this challenge by automatically identifying and retaining only the most significant parameters during training, resulting in simpler, faster, and more accurate models. We demonstrate the practical benefits of the proposed algorithm through two real-world applications: stock market forecasting using the Wilshire index and business profitability prediction based on company financial data. The results show significant improvements over conventional methods: models use fewer parameters–creating simpler, more interpretable solutions–achieve better prediction accuracy, and require less training time. These advantages make the algorithm particularly valuable for business applications where model simplicity, speed, and accuracy are crucial. The method is especially beneficial for organizations with limited computational resources or that require fast model deployment. By automatically selecting the most relevant features, it reduces the need for manual feature engineering and helps practitioners build more efficient predictive models without requiring deep technical expertise in neural network optimization
Early detection of tar spot disease in Zea mays using hyperspectral reflectance and machine learning
Get PDFEnsuring food security and meeting the economic needs of farmers and nations depend heavily on detecting and preventing crop yield losses. Early detection of tar spot caused by Phyllachora maydis is crucial to implementing efficient mitigation actions in the earliest stages of infestation. Currently, visual methods are used for detection, which require extensive training and experience from the operator. However, remote sensing techniques can be used to detect tar spot infestation through the selection of wavelengths present in the maize plant spectral signature. This research proposes using machine learning techniques and logistic regression to determine the first stage of tar spot infestation. The results show that the logistic regression model is the most suitable for detecting this first stage, and the K-Nearest Neighbors Classification and Random Forest Classification algorithms generate the best classification results. This approach can significantly reduce costs in terms of time, labor, and subjective analysis
Improving the transfer learning for batik besurek textile motif classification
Get PDFThis proposed research discussion is a new combination model for classifying batik besurek fabric from the implementation transfer learning with mixed contrast enhancement, activation function, and optimizer method. The size of the batik besurek fabric motif image as an input image is 250×250 with three channels consisting of red, green, and blue totaling five classes, namely kaligrafi, rafflesia, burung kuau, relung paku and rembulan. All images in the dataset will be divided into train data (1540 images), validate data (380 images), and test data (480 images) that are taken directly from the batik store in Bengkulu. The division method used is stratified random sampling to take all the data, shuffles it, and divides the data sets for each class. Based on the experiment results, ResNet50 obtained the best performance compared to MobileNetV2, InceptionV3, and VGG16, with a training accuracy of 99.60%, a validation accuracy of 97.44%, and a testing accuracy of 98.12%. In the improvement experiment phase, the ResNet50 model with Adam optimizer, rectified linear unit (ReLU) activation function and contrast limited adaptive histogram equalization (CLAHE) as the contrast enhancement method obtained the highest test accuracy (98.75%), showing that CLAHE was very effective in improving performance on batik besurek data
Hybrid forecasting methods across varied domains-a systematic review
Get PDFTime series forecasting is one of the links that has developed since early times due to risk management, efficient allocation of resources, performance evaluation, strategic planning, and the formulation of effective policies for individuals, organizations, and societies. Forecasting models have evolved steadily by hybridizing statistical and neural network techniques ensuring efficiency and accurate predictions. In this paper, a systematic review of the literature was made through the preferred reporting items for systematic reviews and meta-analyses (PRISMA) methodology, highlighting the domains that mostly use hybrid techniques by defining the ones with the highest frequency of implementation in each domain we predefined. During the selection process from the 4 selected databases, 2251 works were taken into consideration, of which 25 were the ones that were included in the review process through various filtering steps and exclusion criteria. Ongoing, we defined four main categories where we presented each paper individually by briefly explaining the underlying data, the proposed hybrid forecasting approach and the evaluation performance metrics such as root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). In a summary table, we highlight the most used hybrid methods for each domain, concluding which of the statistical and deep learning methods are mostly applied in the specified domains
Automated vial defect inspection using Gabor wavelets and k-means clustering
Get PDFThis study proposes a machine vision-based defect inspection system for pharmaceutical vials, aiming to ensure the quality and safety of medicinal fluids. The system employs a series of image processing techniques, including denoising, feature extraction using the Gabor wavelet transform, segmentation, clustering with the K-means algorithm, and precise defect identification using the Canny edge operator. Experimental results demonstrate high performance, with recall, precision, accuracy, and F1-score exceeding 98%. Additionally, the proposed method achieves area under the curve-receiver-operating characteristic curve (AUC-ROC) and AUC-precision-recall (PR) values of approximately 98%. The system's average computational time is 355 microseconds, indicating its potential for real-time defect detection. Overall, this approach offers an effective solution for identifying various cosmetic defects such as scratches, bruises, cracks, and black spots, in pharmaceutical vials without the need for vial classification training.
The contribution of artificial intelligence in people with autism: a systematic literature review
Get PDFAutism is a disorder that poses significant challenges in various areas such as health, education, social interaction, and how the world perceives them. The implementation of artificial intelligence in daily life and different fields offers an innovative approach to addressing these challenges, facilitating early detection, support in learning, and social interaction for individuals with this condition. The systematic literature review focuses on studying 50 out of 144 articles obtained from various databases such as EBSCO Host, IEEE Xplore, ScienceDirect, Scopus, ProQuest, and Web of Science. These articles were systematically organized using the preferred reporting items for systematic reviews and meta-analyses (PRISMA) methodology, providing information about machine learning as the most utilized discipline, the types of infrastructure it relies on, and the countries that are at the forefront of this topic. This review will serve as a reference for stakeholders regarding the advancements and contributions of artificial intelligence for individuals with autism