Classification Method Research Articles

Analysis of the retraining strategies for multi-label text message classification in call/contact center systems

Today, in many areas of technology, we can come across applications of various artificial intelligence methods. They usually involve models trained on some specific pool of learning data. Sometimes, however, the data analyzed by these solutions can change its nature over time. This usually results in a decrease in classification efficiency. In such a case, the use of techniques to retrain the originally trained reference models should be considered. One of the industries where the nature of data changes quite dynamically over time is the broadly defined call/contact center systems. An example of a module that is often found in this type of system and that, due to frequently changing marketing campaigns, requires the use of learning techniques is the automatic classification of text data. The paper describes the process of retraining the original reference models used in a multi-label text message classification method dedicated directly to call/contact center systems applications. In order to carry out the retraining process, Polish-language data from the actual archives of a large commercial contact center system and English-language data extracted from a publicly available database were used. The study was conducted for models based on artificial neural networks and bidirectional encoder representations from transformer type models. In addition, two different retraining strategies were studied, the results of which were compared with data obtained from the operation of reference models. As a result of the research work, an improvement of up to 5% in classification efficiency, as described by the metric Emotica was obtained, which means that proper integration of the retraining process brings tangible benefits to the solution tested in the article. Thus, it can also benefit the solutions used in business.

Pulmonary Arterial Anatomical Patterns: a Classification Scheme Based on Lobectomy and 3D-CTBA.

Preoperative evaluation of pulmonary vascular and tracheal routes and variations is of great importance to the surgeon. Three-dimensional computed tomography bronchography and angiography (3D-CTBA) has evolved in recent years with the optimization of 3D reconstruction techniques and artificial intelligence. We aim to apply CT angiography and Exoview 3D reconstruction technology to assess patients' pulmonary arterial tree and its anatomical variants and to try to summarize a set of anatomical typing of the pulmonary arterial tree that is relatively easy and conducive to promoting teaching based on surgical habits of lobectomy. A total of 358 patients hospitalized in the Department of Thoracic Surgery of the First Affiliated Hospital of Soochow University between July 2020 and August 2021 were included in this study. We carefully analyzed the site of emanation, alignment, and number of branches of the pulmonary artery according to a uniform classification method in conjunction with the two-dimensional CT images and transformed them into 3D reconstruction models. Different types of pulmonary artery were observed in 358 cases. We evaluated the complete pulmonary artery tree and counted the number and frequency of major arteries of the pulmonary based on the surgical habits of anatomical lobectomy. The 3D-CTBA technique enables us to adequately assess the anatomy of the pulmonary arteries. Moreover, we provide a practical classification scheme of pulmonary arterial anatomical patterns based on lobectomy and 3D-CTBA. Our data can be used by clinicians in the teaching of pulmonary artery anatomy and the preoperative preparation for anatomical lobectomy.

Comparison between classification algorithms using a multiple correspondence analysis with python.

To analyse the data of a survey questionnaire we apply multiple correspondence analysis “MCA” as a method to help us convert data into a scatter plot, but it is difficult to study it and get good results, so we need to make a classification to facilitate the study. Among the most usable classification methods, hierarchical ascending classification and k-means. To compare them, we conducted a questionnaire on distance studies during the Corona crisis, which included the opinions of 304 university professors from most universities in Algeria; in our application, we used the Python programming language.

PSAEEGNet: pyramid squeeze attention mechanism-based CNN for single-trial EEG classification in RSVP task.

Accurate classification of single-trial electroencephalogram (EEG) is crucial for EEG-based target image recognition in rapid serial visual presentation (RSVP) tasks. P300 is an important component of a single-trial EEG for RSVP tasks. However, single-trial EEG are usually characterized by low signal-to-noise ratio and limited sample sizes. Given these challenges, it is necessary to optimize existing convolutional neural networks (CNNs) to improve the performance of P300 classification. The proposed CNN model called PSAEEGNet, integrates standard convolutional layers, pyramid squeeze attention (PSA) modules, and deep convolutional layers. This approach arises the extraction of temporal and spatial features of the P300 to a finer granularity level. Compared with several existing single-trial EEG classification methods for RSVP tasks, the proposed model shows significantly improved performance. The mean true positive rate for PSAEEGNet is 0.7949, and the mean area under the receiver operating characteristic curve (AUC) is 0.9341 (p < 0.05). These results suggest that the proposed model effectively extracts features from both temporal and spatial dimensions of P300, leading to a more accurate classification of single-trial EEG during RSVP tasks. Therefore, this model has the potential to significantly enhance the performance of target recognition systems based on EEG, contributing to the advancement and practical implementation of target recognition in this field.

Augmented Reality and Artificial Intelligence Medical Waste Classification System and Method

There are four categories of medical waste that cannot be mixed as this will cause serious problems suchas environmental pollution or infection. In the past, the classification of medical waste often involved a lot ofhuman and material resources to process, with workers at risk of exposure to infectious substances. Therefore, an augmented reality (AR) and artificial intelligence (AI) medical waste classification system and method were developed. This innovative medical waste classification system and method combines AR and AI identification technology to reduce the risk of manual judgment errors by clinical staff when handling medical waste.

Common radiographic indices used to measure patellar height do not consistently identify patella alta and lack interchangeability between measurements.

Abnormal patellar height has been identified as a source of aberrant mechanical functioning within the patellofemoral joint. The purpose of this study is to examine the statistical agreement among three commonly used classification methods: Blackburne-Peel (BPI), Caton-Deschamps (CDI) and Insall-Salvati (ISR), by evaluating (1) the rates of patella alta identification and (2) the ability for one index to predict another. One hundred lateral knee radiographs were evaluated using BPI, CDI and ISR to classify each knee as patella normal, patella alta or patella baja. Linear regression analysis was performed to evaluate the relationship between each index. Conversion equations were then derived using the reported linear regression best-fit line, comparing each pair of indices. Patella alta was identified in 15 knees using BPI, 15 using CDI and 25 using ISR. A total of seven knees were classified as patella alta by all BPI, CDI and ISR. Statistical analysis revealed significant correlation (p ≤ 0.001) among BPI and CDI (R2 = 0.706), BPI and ISR (R2 = 0.328) and CDI and ISR (R2 = 0.288). Wilcoxon Signed-Rank test between the three indices revealed no significant difference between the means of converted and original indices. Despite their significant correlations and adequate reproducibility, variability between common patellar height indices render predictions and conversions between BPI, CDI and ISR inequivalent. Users of these indices must be aware of their incongruent properties when considering application to patients in the clinical setting. Furthermore, it remains unclear which patellar height measurement technique is the correct index to use in a given knee. This study highlights the need for further investigation to create a reliable and standardised method for identifying patella height. Level IV.

From citizen science to AI models: Advancing cetacean vocalization automatic detection through multi-annotator campaigns

Continuous underwater Passive Acoustic Monitoring (PAM) has emerged as a strong tool for cetacean research. To handle the vast volume of collected data, it is essential to employ automated detection and classification methods. The recent advancement of deep learning, involving model training and testing, requires a large amount of labeled data. These labels are derived through the manual annotation of audio files often reliant on human experts. Based on an annotation campaign focusing on blue whale calls in the Indian Ocean involving 19 novice annotators and an expert in bioacoustics, this study explores the integration of novice annotators in marine bioacoustics research, through citizen science programs, which could drastically increase the size of labeled datasets and enhance the performance of detection and classification models. The analysis reveals distinctive annotation profiles influenced by the complexity of vocalizations and the annotators' strategies, ranging from conservative to permissive. To address the challenges of annotation discrepancies, Convolutional Neural Networks (CNNs) are trained on annotations from both novices and the expert. The results show variations in model performance. Our work highlights the importance of annotation guidelines encouraging a more conservative approach to improve overall annotation quality. In an effort to optimize the potential of multi-annotation and mitigate the presence of noisy labels, two annotation aggregation methods (majority voting and soft labeling) are proposed and tested. The results demonstrate that both methods, particularly when a sufficient number of annotators are involved, significantly improve model performance and reduce variability: the standard deviation of the area under PR and ROC curves fall under 0.02 for both vocalizations with 13 aggregated annotators, while it was at 0.17 and 0.21 for the Blue Whale Dcalls and 0.05 and 0.04 for the SEIO PBW vocalizations with all annotator separately. Moreover, these aggregation methods enable the training of models using non-expert annotations that achieve performances of models trained with expert annotations. These findings suggest that crowdsourced annotations from novice annotators can be a viable alternative to expert annotations.

Cross-Scene Hyperspectral Image Classification Based on Graph Alignment and Distribution Alignment

A domain alignment-based hyperspectral image (HSI) classification method was designed to address the heterogeneity in resolution and band between the source domain and target domain datasets of cross-scene hyperspectral images, as well as the resulting reduction in common features. Firstly, after preliminary feature extraction, perform two domain alignment operations: image alignment and distribution alignment. Image alignment aims to align hyperspectral images of different bands or time points, ensuring that they are within the same spatial reference framework. Distribution alignment adjusts the distribution of features of samples of different categories in the feature space to reduce the distribution differences of the same type of features between two domains. Secondly, adjust the consistency of the two alignment methods to ensure that the features obtained through different alignment methods exhibit consistency in the feature space, thereby improving the comparability and reliability of the features. In addition, this method considers multiple losses in the model from different perspectives and makes comprehensive adjustments through a unified optimization process to more comprehensively capture and utilize the correlation information between data. Experimental results on Houston 2013 and Houston 2018 datasets can improve the hyperspectral prediction performance between datasets with different resolutions and bands, effectively solving the problems of high cost and limited training samples in HSI labeling and significantly improving cross-scene HSI classification performance.

Enhancing surface drainage mapping in eastern Canada with deep learning applied to LiDAR-derived elevation data.

Agricultural dykelands in Nova Scotia rely heavily on a surface drainage technique called land forming, which is used to alter the topography of fields to improve drainage. The presence of land-formed fields provides useful information to better understand land utilization on these lands vulnerable to rising sea levels. Current field boundaries delineation and classification methods, such as manual digitalization and traditional segmentation techniques, are labour-intensive and often require manual and time-consuming parameter selection. In recent years, deep learning (DL) techniques, including convolutional neural networks and Mask R-CNN, have shown promising results in object recognition, image classification, and segmentation tasks. However, there is a gap in applying these techniques to detecting surface drainage patterns on agricultural fields. This paper develops and tests a Mask R-CNN model for detecting land-formed fields on agricultural dykelands using LiDAR-derived elevation data. Specifically, our approach focuses on identifying groups of pixels as cohesive objects within the imagery, a method that represents a significant advancement over pixel-by-pixel classification techniques. The DL model developed in this study demonstrated a strong overall performance, with a mean Average Precision (mAP) of 0.89 across Intersection over Union (IoU) thresholds from 0.5 to 0.95, indicating its effectiveness in detecting land-formed fields. Results also revealed that 53% of Nova Scotia's dykelands are being used for agricultural purposes and approximately 75% (6924 hectares) of these fields were land-formed. By applying deep learning techniques to LiDAR-derived elevation data, this study offers novel insights into surface drainage mapping, enhancing the capability for precise and efficient agricultural land management in regions vulnerable to environmental changes.

Sound classification with time-frequency features in forest environment

The study of forest sound classification has drawn more attention recently due to its potential for illegal activities and natural disaster monitoring. Based on the forest sound classification dataset (FSC22), a dataset specific to possible sound existing in the forest, five classification methods are utilized to investigate the relationship between recognition accuracy and the number of sound acoustic features, as well as the number of target classes. The results confirmed that extreme random forest is the best method for forest sound classification, with an accuracy of around 70% when the target class number is above 20. Further, Mel-frequency cepstral coefficients are the critical feature for sound classification, while fuzzy labels in the dataset may reduce the success rate of recognition.

MediDRNet: Tackling category imbalance in diabetic retinopathy classification with dual-branch learning and prototypical contrastive learning

Background and ObjectiveThe classification of diabetic retinopathy (DR) aims to utilize the implicit information in images for early diagnosis, to prevent and mitigate the further worsening of the condition. However, existing methods are often limited by the need to operate within large, annotated datasets to show significant advantages. Additionally, the number of samples for different categories within the dataset needs to be evenly distributed, because the characteristic of sample imbalance distribution can lead to an excessive focus on high-frequency disease categories, while neglecting the less common but equally important disease categories. Therefore, there is an urgent need to develop a new classification method that can effectively alleviate the issue of sample distribution imbalance, thereby enhancing the accuracy of diabetic retinopathy classification. Methods:In this work, we propose MediDRNet, a dual-branch network model based on prototypical contrastive learning. This model adopts prototype contrastive learning, creating prototypes for different levels of lesions, ensuring they represent the core features of each lesion level. It classifies by comparing the similarity between data points and their category prototypes. Our dual-branch network structure effectively resolves the issue of category imbalance and improves classification accuracy by emphasizing subtle differences in retinal lesions. Moreover, our approach combines a dual-branch network with specific lesion-level prototypes for core feature representation and incorporates the convolutional block attention module for enhanced lesion feature identification. Results:Our experiments using both the Kaggle and UWF classification datasets have demonstrated that MediDRNet exhibits exceptional performance compared to other advanced models in the industry, especially on the UWF DR classification dataset where it achieved state-of-the-art performance across all metrics. On the Kaggle DR classification dataset, it achieved the highest average classification accuracy (0.6327) and Macro-F1 score (0.6361). Particularly in the classification tasks for minority categories of diabetic retinopathy on the Kaggle dataset (Grades 1, 2, 3, and 4), the model reached high classification accuracies of 58.08%, 55.32%, 69.73%, and 90.21%, respectively. In the ablation study, the MediDRNet model proved to be more effective in feature extraction from diabetic retinal fundus images compared to other feature extraction methods. Conclusion:This study employed prototype contrastive learning and bidirectional branch learning strategies, successfully constructing a grading system for diabetic retinopathy lesions within imbalanced diabetic retinopathy datasets. Through a dual-branch network, the feature learning branch effectively facilitated a smooth transition of features from the grading network to the classification learning branch, accurately identifying minority sample categories. This method not only effectively resolved the issue of sample imbalance but also provided strong support for the precise grading and early diagnosis of diabetic retinopathy in clinical applications, showcasing exceptional performance in handling complex diabetic retinopathy datasets. Moreover, this research significantly improved the efficiency of prevention and management of disease progression in diabetic retinopathy patients within medical practice. We encourage the use and modification of our code, which is publicly accessible on GitHub: https://github.com/ReinforceLove/MediDRNet .

Feature selection for multilabel classification with missing labels via multi-scale fusion fuzzy uncertainty measures

Numerous high-dimension multilabel data are generated, posing a challenge for multilabel learning. Building effective learning models with discriminative features is essential to improve the performance of multilabel learning. Multilabel feature selection can filter out the discriminative features according to their contribution to classification. However, ambiguity, uncertainty, and missing labels coexist in real-life multilabel data, which brings adverse effects to multilabel feature selection. The multi-scale fuzzy rough set gives an effective way to mine intrinsic knowledge hidden in uncertain data. This paper first extends the multi-scale learning to multilabel data with missing labels and proposes a feature selection method for multilabel classification with missing labels via multi-scale fusion fuzzy uncertainty measures called FSMML. The missing label space construction and feature evaluation metric are carefully investigated in the framework of multi-scale learning. A multilabel multi-scale learning strategy is formalized with the fuzzy granularity cognitive mechanism as the core, and the multi-scale fusion fuzzy label learning is given to reconstruct the missing label space. Then, a novel multilabel multi-scale fuzzy rough sets with missing labels is developed, and the significance of each scale is quantified. Moreover, some multi-scale fusion fuzzy uncertainty measures are defined by capturing the sample fuzzy similarity in the feature and reconstructed label spaces. Accordingly, the relevance between features and label set and the interactivity and redundancy between features in feature evaluation are discussed. Finally, FSMML chooses high-quality features to maximize relevance and interactivity and minimize redundancy. Extensive experiments demonstrate the effectiveness of FSMML on fifteen datasets with missing labels.

Multi-Label Noise Robust Collaborative Learning for Remote Sensing Image Classification.

The development of accurate methods for multi-label classification (MLC) of remote sensing (RS) images is one of the most important research topics in RS. The MLC methods based on convolutional neural networks (CNNs) have shown strong performance gains in RS. However, they usually require a high number of reliable training images annotated with multiple land-cover class labels. Collecting such data is time-consuming and costly. To address this problem, the publicly available thematic products, which can include noisy labels, can be used to annotate RS images with zero-labeling cost. However, multi-label noise (which can be associated with wrong and missing label annotations) can distort the learning process of the MLC methods. To address this problem, we propose a novel multi-label noise robust collaborative learning (RCML) method to alleviate the negative effects of multi-label noise during the training phase of a CNN model. RCML identifies, ranks, and excludes noisy multi-labels in RS images based on three main modules: 1) the discrepancy module; 2) the group lasso module; and 3) the swap module. The discrepancy module ensures that the two networks learn diverse features, while producing the same predictions. The task of the group lasso module is to detect the potentially noisy labels assigned to multi-labeled training images, while the swap module is devoted to exchange the ranking information between two networks. Unlike the existing methods that make assumptions about noise distribution, our proposed RCML does not make any prior assumption about the type of noise in the training set. The experiments conducted on two multi-label RS image archives confirm the robustness of the proposed RCML under extreme multi-label noise rates. Our code is publicly available at: https://www.noisy-labels-in-rs.org.

Applying self-supervised learning to network intrusion detection for network flows with graph neural network

Graph Neural Networks (GNNs) have garnered intensive attention for Network Intrusion Detection System (NIDS) due to their suitability for representing the network traffic flows. However, most present GNN-based methods for NIDS are supervised or semi-supervised. Network flows need to be manually annotated as supervisory labels, a process that is time-consuming or even impossible, making NIDS difficult to adapt to potentially complex attacks, especially in large-scale real-world scenarios. The existing GNN-based self-supervised methods focus on the binary classification of network flow as benign or not, and thus fail to reveal the types of attack in practice. This paper studies the application of GNNs to identify the specific types of network flows in an unsupervised manner. We first design an encoder to obtain graph embedding, that introduces the graph attention mechanism and considers the edge information as the only essential factor. Then, a self-supervised method based on graph contrastive learning is proposed. The method samples center nodes, and for each center node, generates subgraph by it and its direct neighbor nodes, and corresponding contrastive subgraph from the interpolated graph, and finally constructs positive and negative samples from subgraphs. Furthermore, a structured contrastive loss function based on edge features and graph local topology is introduced. To the best of our knowledge, it is the first GNN-based self-supervised method for the multiclass classification of network flows in NIDS. Detailed experiments conducted on four real-world datasets (NF-Bot-IoT, NF-Bot-IoT-v2, NF-CSE-CIC-IDS2018, and NF-CSE-CIC-IDS2018-v2) systematically compare our model with the state-of-the-art supervised and self-supervised models, illustrating the considerable potential of our method. Our code is accessible through https://github.com/renj-xu/NEGSC.

Factor annealing decoupling compositional training method for imbalanced hyperspectral image classification

Factor annealing decoupling compositional training method for imbalanced hyperspectral image classification

Russian and Arabic documentary tradition: syntactic aspect

The object of the study is documents of business correspondence in Russian-Arabic official communication; the subject of the study is the syntactic features of Russian and Arabic business documents, the similarities and differences between Russian and Arabic syntax in business communication. Syntactic organization is very important for business writing; it serves as the final stage of its structure. The scientific novelty of the study lies in the identification of universal and culturally determined features of the syntactic structure in relation to Russian and Arabic-language business writing. Among the universal elements are the use of sentences that convey the non-personal nature of communication, indirect expression of imperativeness (using forms of the subjunctive mood, interrogative sentences), introductory words with the semantics of politeness. Original syntactic constructions can be used to achieve a similar communicative effect. The research was carried out on the basis of methods of analysis, synthesis, observation, description. Methods of component analysis, interpretation and classification, a system-structural method, and elements of a functional approach were used. As a result of the study, it was noted that the differences in the syntactic structure of Russian- and Arabic-language business letters are culturally and historically determined. The specifics of Arabic business syntax were influenced by increased requirements for the expression of politeness in society (indication of the regalia of the addressee), the predominance of the collective over the personal in culture, the origin of the official business style from the artistic, and not from the colloquial (long sentences). It is concluded that it is necessary to take into account cultural and historical factors both in the process of constructing business letters and in the process of their mutual translation. Among the important techniques of Russian-Arabic translation are combining sentences and increasing content

A relationship-aware mutual learning method for lightweight skin lesion classification

A relationship-aware mutual learning method for lightweight skin lesion classification

A Fuzzy-Probabilistic Representation Learning Method for Time Series Classification

A Fuzzy-Probabilistic Representation Learning Method for Time Series Classification

Analysis of stream flow and its impact on thermal performance in shell-side of heat exchangers based on flow dead zone

The performance of shell-and-tube heat exchanger (STHE) is primarily determined by different flow paths on the shell-side. The flow path and its impact on the shell-side flow in segmental baffle shell-and-tube heat exchanger with clearance (SC-STHE) is investigated qualitatively and quantitatively, by adopting stream classification and dead zone analysis method. By utilizing Residence Time Distribution (RTD) curves, the influence of different streams on fluid flow and heat transfer is discussed to reveal the intricate interconnections within the shell-side of SC-STHE. Then the cases with baffle-tube clearance (0.2–1.0 mm) and baffle-shell clearance (1.0–5.0 mm) ranges are studied. The study identifies that optimal comprehensive performance is achieved with a baffle-tube clearance of 0.6 mm, resulting in a 3.5 % improvement compared to a clearance of 0.2 mm. An increase of baffle-shell clearance results in a reduction in comprehensive performance by 8.23 % to 19.54 %. The interactions among stream flows and potential mechanisms are discussed by analyzing the influence of each flow path on the flow dead zone volume. It is revealed that the presence of baffle-tube and baffle-shell clearances reduces the volume fraction of flow dead zones compared to structures without clearances. A-stream and E-stream are identified as pivotal, mutually influencing B-stream and C-stream while maintaining relative independence. The result indicates that an increase in baffle-tube clearance results in a 7.60 % expansion of the volume of the dead zone within the B-stream, while in the C-stream, it decreases by 2.58 %. Conversely, an increase in baffle-shell clearance reduces the volume of the dead zone in the B-stream by 5.04 % but increases it in the C-stream by 31.95 %. The flow dead zones are mainly distributed in the B-stream and C-stream regions, and the diversion effect of leakage flows improves the fluid stagnation but also leads to a decrease in turbulence intensity and flow velocity. This study reveals the complex interaction mechanisms among the stream flows, facilitating further analysis and optimization in STHE design.

Memory Shapelet Learning for Early Classification of Streaming Time Series.

Early classification predicts the class of the incoming sequences before it is completely observed. How to quickly classify streaming time series without losing interpretability through early classification method is a challenging problem. A novel memory shapelet learning framework for early classification is proposed in this article. First, a memory distance matrix is introduced to store the historical characteristics of streaming time series, which can alleviate repetitive calculations caused by the growing length of time series. Second, early interpretable shapelets are extracted in the proposed method by optimizing both accuracy objective and earliness objective simultaneously. The proposed method employs end-to-end learning, which allows the model to directly learn early shapelets without the necessity of searching for numerous candidate shapelets. Third, an objective function of memory shapelet learning is proposed by overall considering accuracy and earliness, which can be optimized by gradient descent algorithm. Finally, experiments are conducted on benchmark dataset UCR, Tennessee Eastman process, and real-world aluminum electrolysis process in China. Comparable results with other state-of-the-art methods demonstrate the superior performance of the proposed method in interpretability, accuracy, earliness, and time complexity.