Classification Applications Research Articles

Variability in hot sub-luminous stars and binaries: Machine-learning analysis of Gaia DR3 multi-epoch photometry

Hot sub-luminous stars represent a population of stripped and evolved red giants that is located on the extreme horizontal branch. Since they exhibit a wide range of variability due to pulsations or binary interactions, it is crucial to unveil their intrinsic and extrinsic variability to understand the physical processes of their formation. In the Hertzsprung-Russell diagram, they overlap with interacting binaries such as cataclysmic variables (CVs). By leveraging the most recent clustering algorithm tools, we investigate the variability of 1,576 candidate hot subdwarf variables using comprehensive data from Gaia DR3 multi-epoch photometry and Transiting Exoplanet Survey Satellite (TESS) observations. We present a novel approach that uses the t-distributed stochastic neighbour embedding and the uniform manifold approximation and projection dimensionality reduction algorithms to facilitate the identification and classification of different populations of variable hot subdwarfs and CVs in a large dataset. In addition to the publicly available Gaia time-series statistics table, we adopted additional statistical features that enhanced the performance of the algorithms. The clustering results led to the identification of 85 new hot subdwarf variables based on Gaia and TESS light curves and of 108 new variables based on Gaia light curves alone, including reflection-effect systems, HW Vir, ellipsoidal variables, and high-amplitude pulsating variables. A significant number of known CVs (140) distinctively cluster in the 2D feature space among an additional 152 objects that we consider candidates for new CVs. This study paves the way for more efficient and comprehensive analyses of stellar variability from ground- and space-based observations, and for the application of machine-learning classifications of candidate variable stars in large surveys.

Unlocking the Power of Antioxidant Nanoparticles: Insights into Classification, Formulation, Characterization, and Biomedical Applications

Unlocking the Power of Antioxidant Nanoparticles: Insights into Classification, Formulation, Characterization, and Biomedical Applications

Analysis of the relationship between 18F-FDG PET/CT and ultrasound BI-RADS classification and their combined application in the diagnosis of breast diseases.

To analyze the relationship between fluorine-18-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) and ultrasound breast imaging reporting and data system (BI-RADS) classification, and to evaluate the diagnostic value of their combined application in breast diseases. A retrospective analysis was conducted on the 18F-FDG PET/CT images and ultrasound BI-RADS classification data of 110 patients with suspected breast cancer treated at our hospital from July 2020 to May 2022. Pearson correlation analysis was used to assess the relationship between the maximum standardized uptake value (SUVmax) and BI-RADS classification. Using pathology or long-term follow-up results as the "gold standard," the diagnostic value of 18F-FDG PET/CT, ultrasound BI-RADS classification, and their combined application in breast diseases was analyzed. Based on the "gold standard" of pathology or long-term follow-up, of the 110 patients with suspected breast cancer, 49 were benign, and 61 were malignant. The SUVmax levels of malignant lesions were significantly higher than those of benign lesions (P<0.05). Pearson correlation analysis indicated a low correlation between SUVmax and ultrasound BI-RADS classification (r=0.458, P<0.05). Receiver operating characteristic (ROC) curve analysis showed that the area under the curve (AUC) of the combined application of SUVmax and ultrasound BI-RADS classification was higher than that of either method alone, both for breast tumors and for patients classified as BI-RADS category 3 to 4. The correlation between SUVmax and ultrasound BI-RADS classification is low (r=0.458), indicating that these two methods assess different biological aspects of breast tumors. However, the combined use of SUVmax and BI-RADS classification significantly enhances diagnostic accuracy, particularly for patients with BI-RADS 3 to 4 lesions. Although this combination improves diagnostic efficacy, 18F-FDG PET/CT should not be used as a primary screening tool but rather as a complementary method in specific clinical scenarios where imaging findings are inconclusive or suspicion of malignancy is high.

Satellite Image Time-Series Classification with Inception-Enhanced Temporal Attention Encoder

In this study, we propose a one-branch IncepTAE network to extract local and global hybrid temporal attention simultaneously and congruously for fine-grained satellite image time series (SITS) classification. Transformer and the temporal self-attention mechanism have been the research focus of SITS classification in recent years. However, its effectiveness seems to diminish in the scenario of fine-grained classification among similar categories, for example, different crop types. Theoretically, most of the existing methods focus on only one type of temporal attention, either global attention or local attention, but actually, both of them are required to achieve fine-grained classification. Even though some works adopt two-branch architecture to extract hybrid attention, they usually lack congruity between different types of temporal attention and hinder the expected discriminating ability. Compared with the existing methods, IncepTAE exhibits multiple methodological novelties. Firstly, we insert average/maximum pooling layers into the calculation of multi-head attention to extract hybrid temporal attention. Secondly, IncepTAE adopts one-branch architecture, which reinforces the interaction and congruity of different temporal information. Thirdly, the proposed IncepTAE is more lightweight due to the use of group convolutions. IncepTAE achieves 95.65% and 97.84% overall accuracy on two challenging datasets, TimeSen2Crop and Ghana. The comparative results with existing state-of-the-art methods demonstrate that IncepTAE is able to achieve superior classification performance and faster inference speed, which is conducive to the large-area application of SITS classification.

Enhancing neural network applications in phonetic classification through population-based incremental learning

Significant progress has been achieved in the field of Automatic Speech Recognition (ASR) thanks to the development of powerful algorithms. Research based on simplified biological models has led to the emergence of a new class called Estimation of Distribution Algorithms (EDA), which preserves significant partial solutions. Population-Based Incremental Learning (PBIL) is a technique that combines stochastic search and optimization. It is a statistical approach with evolutionary computation similar to EDAs, aimed at adapting recognition systems based on artificial Neural Networks (NN). Our main contribution lies in the improvement achieved by PBIL, which outperforms Genetic Algorithms (GA) by 16% and Evolutionary Strategies (ES) by 9.93%.

Burned Olive Trees Identification with a Deep Learning Approach in Unmanned Aerial Vehicle Images

Olive tree orchards are suffering from wildfires in many Mediterranean countries. Following a wildfire event, identifying damaged olive trees is crucial for developing effective management and restoration strategies, while rapid damage assessment can support potential compensation for producers. Moreover, the implementation of real-time health monitoring in olive groves allows producers to carry out targeted interventions, reducing production losses and preserving crop health. This research examines the use of deep learning methodologies in true-color images from Unmanned Aerial Vehicles (UAV) to detect damaged trees, including withering and desiccation of branches and leaf scorching. More specifically, the object detection and image classification computer vision techniques area applied and compared. In the object detection approach, the algorithm aims to localize and identify burned/dry and unburned/healthy olive trees, while in the image classification approach, the classifier categorizes an image showing a tree as burned/dry or unburned/healthy. Training data included true color UAV images of olive trees damaged by fire obtained by multiple cameras and multiple flight heights, resulting in various resolutions. For object detection, the Residual Neural Network was used as a backbone in an object detection approach with a Single-Shot Detector. In the image classification application, two approaches were evaluated. In the first approach, a new shallow network was developed, while in the second approach, transfer learning from pre-trained networks was applied. According to the results, the object detection approach managed to identify healthy trees with an average accuracy of 74%, while for trees with drying, the average accuracy was 69%. However, the optimal network identified olive trees (healthy or unhealthy) that the user did not detect during data collection. In the image classification approach, the application of convolutional neural networks achieved significantly better results with an F1-score above 0.94, either in the new network training approach or by applying transfer learning. In conclusion, the use of computer vision techniques in UAV images identified damaged olive trees, while the image classification approach performed significantly better than object detection.

MGNR: A Multi-Granularity Neighbor Relationship and Its Application in KNN Classification and Clustering Methods.

In the real world, data distributions often exhibit multiple granularities. However, the majority of existing neighbor-based machine-learning methods rely on manually setting a single-granularity for neighbor relationships. These methods typically handle each data point using a single-granularity approach, which severely affects their accuracy and efficiency. This paper adopts a dual-pronged approach: it constructs a multi-granularity representation of the data using the granular-ball computing model, thereby boosting the algorithm's time efficiency. It leverages the multi-granularity representation of the data to create tailored, multi-granularity neighborhood relationships for different task scenarios, resulting in improved algorithmic accuracy. The experimental results convincingly demonstrate that the proposed multi-granularity neighbor relationship effectively enhances KNN classification and clustering methods.

Fractal Belief Rényi Divergence With its Applications in Pattern Classification

Fractal Belief Rényi Divergence With its Applications in Pattern Classification

Transforming Industrial Supervision Systems: A Comprehensive Approach Integrating Machine Learning Techniques and Fuzzy Logic

Abstract In addressing the mounting challenges of industrial supervision systems grappling with intricate processes, this study pioneers a transformative paradigm centered on the SCIMAT cement factory. By seamlessly integrating Machine Learning and Fuzzy Logic, the primary aim is to revolutionize real-time control systems, with a keen focus on cement production. SVM integration into the supervision system, coupled with connectivity to a Programmable Logic Controller (PLC), is complemented by fuzzy real-time controllers’ regression analysis. Rigorous testing and evaluation validate the proposed approach’s reliability, showcasing its effectiveness in discerning optimal system functioning. The system’s practical application within a PLC environment underscores its prowess in issuing commands to industrial equipment, thereby enhancing operational efficiency. Going beyond conventional methodologies, our approach amalgamates SVM classification, fuzzy controllers, and real-time regression analysis, delivering a multifaceted solution for industrial supervision. The system’s standout achievement is an SVM classification accuracy surpassing 94% compared to other classifiers. The K-Nearest Neighbors (K-NN) model demonstrated an accuracy rate of approximately 93.83%. The decision tree model attained an accuracy of around 83.73%. The logistic regression model achieved an accuracy of 80.25%. These models are not only adept at distinguishing optimal functioning from faults but also adept at preserving the linguistic language used by operators. The study’s novelty lies in the holistic integration of SVM and Fuzzy Logic, offering a practical and adaptable solution that not only advances classification accuracy but also significantly reduces maintenance costs, marking a substantial improvement over the traditional methods. This transformative model, validated through SVM classification and practical application, establishes a new standard for flexibility, cost reduction, and overall productivity enhancement in industrial processes.

Application of one-class classification using deep learning technique improves the classification of subvisible particles

Capturing subvisible particles using flow imaging microscopy is useful for evaluating protein aggregates that may induce immunogenicity. Automated labeling is desirable to distinguish harmless components such as silicone oil (SO) from subvisible particles. The one-class classifier, which requires only target class data for model establishment, is suitable for machine learning and proposes a useful solution for distinguishing a subject with heterogeneous but stable distributions, such as SO. However, the effectiveness of the application of one-class classifiers to subvisible particles remains unclear. In this study, we investigated whether deep learning techniques can improve the performance on a variety of images. We prepared datasets using SO and two types of protein aggregates: immunoglobulin G-derived aggregates (AggIgG) and albumin-derived aggregates (AggAlb). The deep-learning technique improved the classification scores for both AggIgG and AggAlb. The classification scores for AggIgG were more satisfactory than those for AggAlb. Cluster analysis revealed that one-class classification using deep learning techniques achieved excellent effectiveness across almost all clusters in classifying AggIgG. Collectively, the deep learning technique remarkably improved the one-class classification of subvisible particles of AggIgG and AggAlb. Combined with deep learning, one-class classification can contribute to the evaluation of subvisible particles, particularly for AggIgG.

Application of digital classification based on AO classification in floating knee injury

ObjectiveA digital classification system based on AO classification was adopted to study and summarize different types of floating knee injuries.MethodsOne hundred cases with a floating knee injury from March 2005 to December 2020 were collected, including 74 males and 26 females with an average age of 43.74 years. The distribution of the affected limbs in there were 51 cases (51.0%) on the left and 46 cases (46.0%) on the right, as well as 3 cases (3.0%) that got bilaterally affected limbs. The femora and tibiae in this system were divided into three parts separately based on the AO classification, which was finally composed of six types after pairwise combination. The images in this study would be analyzed by using the classification to summarize the distribution rules of the classification in a floating knee injury.ResultsThe proportion of I-VI type injury was 42.7%(44/103) in type I, 18.4%(19/103) in type II, 9.7%(10/103) in type III, 10.7%(11/103) in type IV, 11.7%(12/103) in type V, and 6.8%(7/103) in type VI.ConclusionsThe digital classification system of floating knee injury based on AO classification could classify most cases of floating knee injury and cases with open injury, which referred to a reference effectively for the selection of surgical treatment plan, and would be conducive to the future clinical classification.

A Review of Research on Image Classification Methods

As one of the important research directions in the field of computer vision, image classification has a wide range of applications. Image classification is an important foundation for pattern recognition, machine learning, and artificial intelligence. Image classification generally includes three steps: region of interest selection, feature extraction, and classifier modeling. Among them, feature extraction of images is an important foundation for completing other tasks. In most pattern recognition scenarios, appropriate feature representation is a crucial step that directly affects the performance of the entire classification system. Among them, the most representative is the deep learning method that directly learns complex feature expressions from massive data. This article first reviews the research background and development history of image classification, then analyzes the application of image classification in different fields and lists case studies. It also outlines traditional image classification methods and mainstream image classification techniques in deep learning, including convolutional neural networks, model pre-training, and data augmentation. At the end of the article, the main challenges currently faced by image classification were analyzed, and future directions for improvement and technologies that can be combined were discussed. This review aims to provide researchers in image classification with certain research directions.

Automatic Text Classification With Large Language Models: A Review of openai for Zero- and Few-Shot Classification

While natural language documents, such as intervention transcripts and participant writing samples, can provide highly nuanced insights into educational and psychological constructs, researchers often find these materials difficult and expensive to analyze. Recent developments in machine learning, however, have allowed social scientists to harness the power of artificial intelligence for complex data categorization tasks. One approach, supervised learning, supports high-performance categorization yet still requires a large, hand-labeled training corpus, which can be costly. An alternative approach—zero- and few-shot classification with pretrained large language models—offers a cheaper, compelling alternative. This article considers the application of zero-shot and few-shot classification in educational research. We provide an overview of large language models, a step-by-step tutorial on using the Python openai package for zero-shot and few-shot classification, and a discussion of relevant research considerations for social scientists.

A salient feature establishment tactic for cassava disease recognition

Accurate classification of cassava disease, particularly in field scenarios, relies on object semantic localization to identify and precisely locate specific objects within an image based on their semantic meaning, thereby enabling targeted classification while suppressing irrelevant noise and focusing on key semantic features. The advancement of deep convolutional neural networks (CNNs) paved the way for identifying cassava diseases by leveraging salient semantic features and promising high returns. This study proposes an approach that incorporates three innovative elements to refine feature representation for cassava disease classification. First, a mutualattention method is introduced to highlight semantic features and suppress irrelevant background features in the feature maps. Second, instance batch normalization (IBN) was employed after the residual unit to construct salient semantic features using the mutualattention method, representing high-quality semantic features in the foreground. Finally, the RSigELUD activation method replaced the conventional ReLU activation, enhancing the nonlinear mapping capacity of the proposed neural network and further improving fine-grained leaf disease classification performance. This approach significantly aided in distinguishing subtle disease manifestations in cassava leaves. The proposed neural network, MAIRNet-101 (Mutualattention IBN RSigELUD Neural Network), achieved an accuracy of 95.30 % and an F1-score of 0.9531, outperforming EfficientNet-B5 and RepVGG-B3g4. To evaluate the generalization capability of MAIRNet, the FGVC-Aircraft dataset was used to train MAIRNet-50, which achieved an accuracy of 83.64 %. These results suggest that the proposed algorithm is well suited for cassava leaf disease classification applications and offers a robust solution for advancing agricultural technology.

Identification of Cactus Species Using the MobileNet Convolutional Neural Network Model

The objective of this design is to create a system capable of accurately and precisely classifying different types of succulent cacti. The system design aims to create a succulent cactus classification application using a Convolutional Neural Network (CNN) with the MobileNet architecture. The process involves collecting cactus images, dividing them into training and test datasets, and developing a CNN model to recognize patterns in succulent cactus species. The training data is used to train the CNN model, and the test data evaluates the model's accuracy. The trained model, stored in Tf.lite format, successfully classifies 15 cactus types, achieving high accuracy by employing preprocessing steps like resizing, normalization, and background removal. Over 1,200 cactus images were taken with a smartphone, categorized into 15 classes, and prepared to ensure optimal lighting, angle, background, and resolution (224x224 pixels). The MobileNet model was chosen for its high accuracy and efficiency. Hardware used includes a Samsung A54 smartphone and an Intel i7 laptop, with software such as Python, Kotlin, and Android Studio facilitating development. This design ensures the application’s accessibility, making it a valuable tool for cactus enthusiasts and the general public to easily identify different succulent cactus types. Testing of the cactus species classification program using the Convolutional Neural Network (CNN) method with the MobileNetV2 architecture yielded strong results, achieving over 90% accuracy in classifying 15 cactus species. The highest training accuracy of 0.9837 was achieved at 150 epochs without early stopping, outperforming other epoch configurations. The model successfully classified species across five main genera—Kalanchoe, Crassula, Echeveria, Haworthia, and Euphorbia. This high accuracy highlights the model's effectiveness, making it a useful tool for cactus enthusiasts and the public to accurately identify and distinguish cactus species.

A study of machine learning applications in healthcare

Abstract. Machine learning, a sub-field of artificial intelligence, has numerous algorithms with applications in image recognition and classification, natural language processing, data prediction, medical diagnosis, and more, achieving significant results. Nowadays, the application of many algorithms of machine learning in the medical field has become a research hotspot and attracted wide attention from society, which plays a key role in disease screening, disease prediction, and assisted diagnosis of diseases. In this paper, we have developed a new approach to the medical field that can be applied in the healthcare industry by using technologies such as k-nearest neighbor (KNN), recurrent neural network (RNN), plain Bayesian (NB), decision tree (DTT), random forest (RF), deep convolutional adversarial network (DCAN), generative adversarial network (GAN), deep convolutional generative adversarial network (DCGAN),deep convolutional neural network (CNN), and support vector machine (SVM). By referring to the existing related literature and application practice results, this paper provides a systematic introduction to the application and research of machine learning in the medical field. This paper also analyses and describes the challenges and difficulties in the implementation of machine learning in the medical field.

Enhancing Image Classification with Super-Resolution Techniques: A Comparative Study of SRResNet and SRGAN

Abstract. This paper explores the use of super-resolution (SR) approaches to enhance the performance of image classification, with particular attention to the effects of various SR models on classification accuracy. Using the ImageNet Dogs dataset for assessment, the paper examines two SR techniques: Super Resolution Generative Adversarial Networks (SRGAN) and Super Resolution Residual Networks (SRResNet). The research demonstrates that deep learning-based SR methods can enhance classification accuracy. Notably, SRResNet is identified as the superior method for improving classification performance, despite generating less visually appealing images compared to SRGAN. These finding highlights that while Generative Adversarial Networks (GANs) and perceptual loss functions can enhance image quality, their impact on classification accuracy may not always be substantial. The study offers important new perspectives on the relative merits of different SR approaches, highlighting the necessity of choosing the right SR methods in accordance with the particular demands of picture classification tasks. The results suggest that SRResNet, with its focus on accuracy over visual appeal, is more effective for boosting classification model performance, offering guidance for optimizing SR methods in practical image classification applications.

Autism, intelligence, language, and adaptive behavior, disentangling a complex relationship.

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder characterized by a range of intellectual and language abilities. Its heterogeneity is acknowledged in modern diagnostics, complicating research and necessitating precision medicine and a multidimensional approach for individualized treatment and accurate assessment. Intellectual and language functioning influence adaptive skills and symptomatology. Thus, assessing adaptive functioning in a multidimensional and multi-informant manner is crucial, highlighting the importance of comprehensive evaluations. This study explores the interplay between autistic traits, demographic variables, IQ, adaptive functioning, and the applicability of ICD-11 classifications. We analyzed data from the initial global evaluation of 60 diverse autistic children (aged 35 to 120 months; IQ range 16 to 118). Parent-reports using the Vineland Adaptive Behavior Scale (VABS-II) were compared with standardized assessments from the PsychoEducational Profile (PEP-3). Children's intellectual levels were assessed using Griffiths Scales of Child Development (Griffiths III) and the Autism Diagnostic Observation Schedule (ADOS-2) was used for autistic traits. They were further classified according to the ICD-11 diagnostic system, functional language, and intellectual functioning levels. Correlations among variables, group comparisons, and multivariate analyses were performed. The analysis indicates a linear effect of IQ on all adaptive scales and the impact of autistic traits on directly measured adaptive functioning. A factorial effect was observed due to changes concerning specific age, intellectual, and linguistic levels, which do not completely align with ICD-11 categorization. Additionally, a negative correlation between intelligence and measured autistic traits was found. Parental age, education level, and age at childbirth were also found to affect various adaptive scales. The study questions the ICD-11's proposed distinctions in IQ and language functioning for ASD, advocating for more refined categorization and developmental considerations. It underscores the intricate relationship between autistic traits, IQ, and communication skills, casting doubt on the precision of diagnostic tools across the spectrum. Parental reports and direct assessments are essential for comprehensive evaluation, with parental education and age influencing children's behaviors and skills. The study calls for a nuanced approach to ASD assessment, integrating various metrics and sources of information for a detailed phenotyping necessary for personalized interventions and biological research.

Evidential Deep Learning: Enhancing Predictive Uncertainty Estimation for Earth System Science Applications

Abstract Robust quantification of predictive uncertainty is a critical addition needed for machine learning applied to weather and climate problems to improve understanding of what is driving prediction sensitivity. Ensembles of machine learning models provide predictive uncertainty estimates in a conceptually simple way but require multiple models for training and prediction, increasing computational cost and latency. Parametric deep learning can estimate uncertainty with one model by predicting the parameters of a probability distribution but does not account for epistemic uncertainty. Evidential deep learning, a technique that extends parametric deep learning to higher-order distributions, can account for both aleatoric and epistemic uncertainty with one model. This study compares the uncertainty derived from evidential neural networks to that obtained from ensembles. Through applications of classification of winter precipitation type and regression of surface layer fluxes, we show evidential deep learning models attaining predictive accuracy rivaling standard methods, while robustly quantifying both sources of uncertainty. We evaluate the uncertainty in terms of how well the predictions are calibrated and how well the uncertainty correlates with prediction error. Analyses of uncertainty in the context of the inputs reveal sensitivities to underlying meteorological processes, facilitating interpretation of the models. The conceptual simplicity, interpretability, and computational efficiency of evidential neural networks make them highly extensible, offering a promising approach for reliable and practical uncertainty quantification in Earth system science modeling. In order to encourage broader adoption of evidential deep learning, we have developed a new Python package, MILES-GUESS (https://github.com/ai2es/miles-guess), that enables users to train and evaluate both evidential and ensemble deep learning.

SimCDL: A Simple Framework for Contrastive Dictionary Learning

In this paper, we propose a novel approach to the dictionary learning (DL) initialization problem, leveraging the SimCLR framework from deep learning in a self-supervised manner. Dictionary learning seeks to represent signals as sparse combinations of dictionary atoms, but effective initialization remains challenging. By applying contrastive learning, we encourage similar representations for augmented versions of the same sample while distinguishing between different samples. This results in a more diverse and incoherent set of atoms, which enhances the performance of DL applications in classification and anomaly detection tasks. Our experiments across several benchmark datasets demonstrate the effectiveness of our method for improving dictionary learning initialization and its subsequent impact on performance in various applications.