IAES International Journal of Artificial Intelligence (IJ-AI)
Not a member yet
1769 research outputs found
Sort by
Graph-based methods for transaction databases: a comparative study
Get PDFThere has been an increased demand for structured data mining. Graphs are among the most extensively researched data structures in discrete mathematics and computer science. Thus, it should come as no surprise that graph-based data mining has gained popularity in recent years. Graph-based methods for a transaction database are necessary to transform all the information into a graph form to conveniently extract more valuable information to improve the decision-making process. Graph-based data mining can reveal and measure process insights in a detailed structural comparison strategy that is ready for further analysis without the loss of significant details. This paper analyzes the similarities and differences among four of the most popular graph-based methods that is applied to mine rules from transaction databases by abstracting them out as a concrete high-level interface and connecting them into a common space
Electroencephalogram denoising using discrete wavelet transform and adaptive noise cancellation based on information theory
Get PDFOne of the most frequently used techniques for removing background noise from electroencephalogram (EEG) data is adaptive noise cancellation (ANC). Nonetheless, there exist two primary disadvantages associated with the adaptive noise reduction of EEG signals: the adaptive filter, which is supposed to be an approximation of contaminated noise, lacks the reference signal. The mean squared error (MSE) criterion is frequently employed to achieve this goal in adaptive filters. The MSE criterion, which only considers second-order errors, cannot be used since neither the EEG signal nor the EOG artifact are Gaussian. In this work, we employ an ANC system, deriving an estimate of EOG noise with a discrete wavelet transform (DWT) and input this signal into the reference of the ANC system. The entropy-based error metric is used to reduce the error signal instead of the MSE. Results from computer simulations demonstrate that the suggested system outperforms competing methods with respect to root-mean-square-error, signal-to-noise ratio, and coherence measurements
Optimizing the gallstone detection process with feature selection statistical analysis algorithm
Get PDFEarly detection is one form of early anticipation in treating gallstone disease patients using medical images. However, the problem that exists is that there are still many shortcomings in medical images, such as noise in the image that causes the detection process to not run optimally. Based on this, this study aims to carry out the process of detecting gallstone objects in magnetic resonance cholangiopancreatography (MRCP) images by optimizing the performance of extraction techniques for feature selection. Optimization of extraction techniques in feature selection is carried out using the performance of the feature selection statistics analysis (FSSA) algorithm. The performance of the FSSA algorithm can provide improvements in the feature selection process by excelling in the performance of classification methods such as k-nearest neighbor (KNN), support vector machine (SVM), and artificial neural network (ANN), and the Pearson correlation (PC) method. Based on the tests that have been carried out, the performance of the FSSA algorithm in the detection process provides an accuracy level of 95.69%, a sensitivity of 89.65%, and a specificity of 98.43%. Overall, this study can contribute to the development of extraction and provide a significant technical impact on optimizing the gallstone detection process
A page rank-based analytical design of effective search engine optimization
Get PDFSearch engine optimization (SEO) is an important internet marketing strategy and process that facilitates maximizing an intended website’s visibility with search engine results. It is widely employed nowadays to improve traffic volume or quality from search engines to a particular website. Even though a significant number of publications imply the essential aspects of SEO, only a few provide generalized ideas to deal with the complex structure of the web. Also, the critical issues of content quality, site popularity, keyword density, and publicity factors were not much considered in the traditional ranking algorithms during SEO processes. This has negatively influenced the retrieval rate in the existing SEO techniques, and consequently, inadequate search results were obtained through search engines. Hence, the study considers web page ranking as a theoretical basis for the research and addresses these limitations in the existing system. It further improves SEO performance by introducing a unique web-page ranking strategic design to gain higher page rank results. The results of the investigational study show that the proposed system effectively contributes towards SEO with an improved page ranking strategy and also provides higher accuracy in calculating the importance score of web pages which is comparable with popular ranking algorithms such as hyperlink-induced topic search (HITS) and PageRank
High body temperature detection solution through touchless machine for health monitoring
Get PDFThe demand for reliable health monitoring systems has surged in today's health-conscious society. Body temperature monitoring is crucial for preserving health and preventing infectious disease outbreaks. In this study an Arduino uno hardware board with a touchless temperature sensor is proposed to detect elevated body temperature, indicating fever and early signs of illness. The system prioritizes real-time health surveillance, accessibility, and usability, blending seamlessly with normal life. Arduino's versatility allows the system to function covertly, uphold privacy and autonomy, and foster wellbeing. The goal is to highlight the system's ability to function covertly, uphold privacy and autonomy, and foster wellbeing. This technology exemplifies the synergy between personal wellness and contemporary technologies, offering a useful and adaptable fever detection solution for various contexts, including homes and public areas
Crop classification using object-oriented method and Google Earth Engine
Get PDFAgriculture crop monitoring in real-time is crucial in formulating effective agricultural practices and management policies. The primary goal of the investigation is to explore how the utilization of Sentinel-1 data and its fusion with Sentinel-2 impact crop classification accuracy in a fragmented agricultural landscape in the Yavatmal District of Maharashtra, India. Pixel based classification and object-oriented classification approaches were implemented on Google Earth Engine (GEE), and obtained results were compared for different combinations of optical and microwave features. The research revealed that the object-based technique performed better than the pixel-based approach, with a 3.5% increase in overall accuracy. For 2022, crop-type mapping was generated with overall accuracies varying from 85.5% to 61% and a kappa coefficient between 0.77 and 0.37. These overall accuracies imply that joint use of optical and radar data has given a 24% improvement in overall accuracy compared to use of only optical data. In addition, the temporal change in the backscatter coefficients and different vegetation indices for different crops were examined over crop growth cycle. This work demonstrates the fusion of Sentinel-1 and Sentinel-2 data to classify wheat, chickpea, other crops, water and urban areas
Co-training pseudo-labeling for text classification with support vector machine and long short-term memory
Get PDFThe scarcity of labeled data may hamper training text-processing models. In response to this issue, a novel and intriguing strategy that combines the co-training method and pseudo-labeling design is applied to enhance the model's performance. This method, a component of an efficient semi-supervised learning paradigm for processing and comprehending text, is a fresh perspective in the field. The model, which combines a support vector machine (SVM) for classification and long short-term memory (LSTM) for text sequence interpretation, is a unique approach. By introducing samples that may be marginalized in the labeled data, the co-training approach could help solve the class imbalance problem by using a small amount of labeled data and the rest unlabeled. This study assesses the model's performance using a student dataset from higher education institutions to establish a threshold for each model's degree of confidence and ascertain how much the model can be generalized depending on the threshold. The SVM threshold was calculated as >=0.88, and the LSTM threshold was calculated as >=0.5 using a mixture of confidence metrics
Enhancing face mask detection performance with comprehensive dataset and YOLOv8
Get PDFIn the context of the COVID-19 pandemic and the risk of similar infectious diseases, monitoring and promoting public health measures like wearing face masks have become crucial in controlling virus transmission. Deep learning-based mask recognition systems play an important role, but their effectiveness depends on the quality and diversity of training datasets. This study proposes the diverse and robust dataset for face mask detection (DRFMD), designed to address limitations of existing datasets and enhance mask recognition models' performance. DRFMD integrates data from sources such as AIZOO, face mask detector by Karan-Malik (KFMD), masked faces (MAFA), MOXA3K, properly wearing masked face detection dataset (PWMFD), and the Zalo AI challenge 2022, comprising 14,727 images with 29,846 instances, divided into training, validation, and testing sets. The dataset's scale and diversity ensure higher accuracy and better generalization for mask recognition models. Experiments with variations of the YOLOv8 model (n, s, m, l, x), an advanced object detection algorithm, on the DRFMD dataset, demonstrate superior performance through metrics like precision, recall, and mAP@50. Additionally, comparisons with previous dataset like FMMD show that models trained on DRFMD maintain strong generalization capabilities and higher performance. This study significantly contributes to improving accuracy of public health monitoring systems, aiding in the prevention of hazards from infectious diseases and air pollution
Comparison among search algorithms for hyperparameter of support vector machine optimization
Get PDFSupport vector machine (SVM) is widely used in machine learning for classification and regression tasks, but its performance is highly dependent on hyperparameter tuning. Therefore, fine-tuning these parameters is key to improving accuracy and generality. Recently, many researchers have focused only on applying different algorithms to optimize these parameters. There is a shortage of studies that compare the performance of these methods. Hence, research is needed to compare the performance of these algorithms for the hyperparameters of the SVM optimization problem. This paper compares five optimization algorithms for tuning SVM hyperparameters: grid search (GS), random search (RS), Bayesian optimization (BO), genetic algorithm (GA), and the novel chemical reaction optimization (CRO) algorithm. Experimental results on benchmark datasets such as iris, digits, wine, breast cancer Wisconsin, and credit card fraud demonstrate that CRO consistently outperforms other methods in terms of classification scoring metrics and computational time. It achieves improvements in accuracy, precision, recall, and F1-score of up to 1% on balanced datasets and up to 10% on highly imbalanced datasets such as credit card fraud. It also reduces computation time by up to 50% compared to GS, BO, and RS. These findings suggest that CRO is a promising approach for hyperparameter optimization (HPO) of SVM models
Optimizing firewall timing for brute force mitigation with random forests
Get PDFMitigating brute force attacks remains a critical challenge in cybersecurity, requiring intelligent and adaptive solutions. This research introduces an approach to optimizing firewall deployment timing for enhanced brute force mitigation using pattern recognition techniques with the random forest algorithm. Leveraging the UNSW-NB15 dataset, comprehensive preprocessing and exploratory data analysis (EDA) were performed to ensure the dataset's suitability for machine learning applications. The study utilized a structured workflow, splitting the dataset into training and testing subsets to rigorously evaluate the model's performance. The proposed random forest model achieved a high accuracy of 98.87%, supported by precision, recall, and F1-scores that confirm its effectiveness in distinguishing normal and attack traffic. The confusion matrix further validated the model’s robustness, highlighting its potential in improving the efficiency of firewall deployment. These findings demonstrate the critical role of advanced machine learning techniques in enhancing cybersecurity defenses, particularly in mitigating brute force attacks through optimized, data-driven strategies