检索结果-内蒙古大学图书馆

Robust Network Traffic classification Based on Information Bottleneck Neural Network

IEEE ACCESS 2024年 12卷 150169-150179页

作者： Lin, Wei Chen, Yu Fuzhou Inst Technol Sch Comp & Informat Sci Fuzhou 350506 Peoples R China

Network Traffic classification (NTC) plays a critical role in modern network security and management, used in various applications such as performance monitoring, anomaly detection, and bandwidth management. Traditional methods like port-based identification and deep packet inspection (DPI) face limitations due to evolving encryption techniques and dynamic port usage. Recent advances in machine learning have improved NTC but often struggle with data labeling, encrypted traffic, and model complexity. This paper introduces a novel Information Bottleneck Neural Network (IBNN) to address these challenges by efficiently extracting essential features from raw traffic data while minimizing irrelevant information. The IBNN framework balances model complexity and discriminative feature preservation, leading to enhanced classification accuracy and robustness, particularly for encrypted traffic. Through information-centric feature extraction and relevant information compression, the proposed method outperforms existing techniques about 5-7% in terms of accuracy, as well as F1 score on both ISCX 2012 and UNB CIS datasets, delivering superior resilience and adaptability across diverse network scenarios. Extensive experiments demonstrate the efficacy of IBNN, positioning it as a significant advancement in the field of NTC.

关键词： Telecommunication traffic Deep learning Neural networks Feature extraction Noise Resilience Cryptography Training Quality of service Protocols classification algorithms Artificial neural networks Communication networks Network traffic classification artificial neural network information bottleneck communication networks

来源：评论

学校读者我要写书评

暂无评论

Novel classification Method to Predict the Accuracy of UWB Ranging Estimates

引用

IEEE ACCESS 2024年 12卷 33659-33670页

作者： Arsuaga, Meritxell Ochoa-De-Eribe-Landaberea, Aitor Zamora-Cadenas, Leticia Arrizabalaga, Saioa Velez, Igone Ibermat Ayesa Co Donostia San Sebastian 20009 Spain CEIT Basque Res & Technol Alliance BRTA Donostia San Sebastian 20018 Spain Univ Navarra Tecnun Sch Engn Donostia San Sebastian 20018 Spain

Real time location systems (RTLSs) are becoming more relevant in a more data driven economy and society due to their wide range of application cases. When the location of an object needs to be tracked with high accuracy, ultra wideband (UWB) technology is usually the best option. Nevertheless, UWB ranging estimates are not completely immune to some sources of error such as non line of sight (NLOS) or multipath conditions. Thus, this paper proposes a real-time classification model based on machine learning (ML) to predict if received ranging estimates are in line of sight (LOS) or NLOS conditions and discard those in NLOS. However, it is also shown that classifying measurements as LOS or NLOS does not guarantee detecting inaccurate ranging estimates, since LOS measurements can also yield large errors. As an example, the ranging root mean square error (RMSE) of the data labelled as LOS in a UWB based localization system database in the literature is of 0.714 m, significantly higher than the theoretical accuracy of a UWB system. Thus, a novel ML-based classification model is proposed to predict the magnitude of the ranging error. After applying the proposed classification model in the same data, the ranging RMSE of those ranging samples classified as most accurate is of only 0.183 m, significantly lower than the best RMSE we can obtain on the classical LOS/NLOS classification approach.

关键词： Distance measurement classification algorithms Real-time systems Sensors Time measurement Ultra wideband technology Proposals Machine learning Random forests Location awareness Data models Predictive models DWM1000 machine learning random forest ranging errors real time location system ultra wideband

来源：评论

学校读者我要写书评

暂无评论

EDLRDPL_Net: A New Deep Dictionary Learning Network for SAR Image classification

引用

IEEE SIGNAL PROCESSING LETTERS 2024年 31卷 1459-1463页

作者： Gao, Xizhan Wei, Kang Niu, Sijie Zhao, Hui Dong, Jiwen Univ Jinan Sch Informat Sci & Engn Shandong Prov Key Lab Network based Intelligent Co Jinan 250022 Peoples R China

SAR image classification is one of the research hotspots in the field of remote sensing. The performance of SAR image classification greatly depends on the learning of features and the design of classifiers. However, traditional SAR image classification methods either focus on learning (deep) features or focus on designing discriminative classifiers, while ignoring the correlation between them, resulting in the learned features and classifiers often not matching. To address the above issues, in this letter a novel classifier, termed entropy based low-rank dictionary pair learning (ELRDPL) method is first designed, which introduces the entropy theory and low-rank constraints into the objective function of dictionary learning for increasing the discriminative ability and decreasing the space occupation. Based on the constructed classifier, an entropy based deep low-rank dictionary pair learning network (EDLRDPL_Net) is then proposed, which performs joint learning of deep features and discriminative dictionaries by embedding the ELRDPL classifier into the deep neural network. Extensive experimental results on four SAR image classification datasets demonstrate the effectiveness of EDLRDPL_Net.

关键词： Dictionaries Entropy Machine learning Radar polarimetry classification algorithms Optimization Image reconstruction Deep dictionary learning discriminative dictionary entropy constraint joint learning network SAR image classification

来源：评论

学校读者我要写书评

暂无评论

Mixed-Embeddings and Deep Learning Ensemble for DGA classification With Limited Training Data

引用

IEEE ACCESS 2025年 13卷 81167-81187页

作者： Morbidoni, Christian Cucchiarelli, Alessandro Spalazzi, Luca Univ G dAnnunzio Dept Business Adm I-65127 Pescara Italy Univ Politecn Marche Dept Informat Engn I-60121 Ancona Italy

Recent papers in the cybersecurity research field of Domain Generation algorithms (DGAs) detection show the increase of performances associated with the introduction of unsupervised neural vectorized representation of domain names in the supervised classification process. In this paper we explore the effectiveness of this approach by proposing a novel mixed pre-trained neural embeddings model which integrates different vectorized representations of domain names: n-grams streams and words. We used the embeddings with two different classifiers, both based on ensemble architectures: a stacking model and an end-to-end multi-input neural architecture. We trained and tested the classifiers with two datasets, differing both in the distribution of domain names between real and DGAs and in the number and type of DGAs. The obtained results show that our solution provides considerable advantages with respect to state-of-the-art single classifiers both in classification accuracy and in the detection of challenging DGAs, such as those based on word dictionaries. The improvement of performance is significant in a particularly relevant operating condition, known as few-shot-learning, where only few examples of DGA-generated domain names are available for the classifier training.

关键词： Training Few shot learning classification algorithms Botnet Telecommunication traffic Long short term memory Training data Malware Domain Name System Transformers Domain generation algorithms (DGA) botnet deep learning LSTM n-grams pre-trained embeddings few-shot learning

来源：评论

学校读者我要写书评

暂无评论

A Predictive Model for the Early Identification of Student Dropout Using Data classification, Clustering, and Association Methods

引用

IEEE REVISTA IBEROAMERICANA DE TECNOLOGIAS DEL APRENDIZAJE-IEEE RITA 2025年 20卷 12-21页

作者： Chicon, Patricia Mariotto Mozzaquatro Paschoal, Leo Natan Sawicki, Sandro Roos-Frantz, Fabricia Frantz, Rafael Z. Univ Cruz Alta UNICRUZ Distance Educ Ctr Cruz Alta Rio Grande Do S Brazil Paulista Univ UNIP Inst Exact Sci & Technol Araraquara SP Brazil Reg Univ Northwestern Rio do Grande UNIJUI Dept Exact Sci & Engn Ijui Rio Grande do S Brazil

Technology development has led to increased data generated in education, sparking interest in information extraction to support educational management through automated data analysis. Using such data to create models identifying students likely to drop out has drawn research interest. A crucial factor in reducing dropout rates is the systematic and early identification of the level of student engagement, especially by detecting the students' behavior profile in the virtual environment, such as grades in assessments. There are predictive models based on data mining processes that identify students prone to dropping out. Unfortunately, the predictive models do not characterize the profiles of these students or the specific trends associated with these profiles. This article aims to fill a gap by presenting a study that identifies and tracks the profiles of undergraduate students likely to drop out, starting with an analysis of academic performance. We propose a predictive model beyond classification by combining data mining techniques such as decision trees, clustering, and frequent pattern analysis. Decision trees, a data mining technique that uses a tree-like graph to represent decisions and their possible consequences, identify students at risk of failure from the entire dataset. Clustering analysis, a data mining technique that groups similar data points together, groups students based on similar characteristics (e.g., students who scored between 0 and 30 points on a specific activity). Frequent pattern analysis, a data mining technique that identifies patterns that occur frequently in a dataset, uncovers the underlying factors contributing to low performance (e.g., identify which activities had the most significant influence on a specific group's low performance). This integrated approach predicts dropout risk with 93.9% precision and provides a deeper understanding of student profiles and the trends associated with academic failure. The model's practical applica

关键词： Predictive models Data models Data mining Random forests Prediction algorithms Decision trees Analytical models Market research Clustering algorithms classification algorithms Data analysis data mining performance prediction

来源：评论

学校读者我要写书评

暂无评论

A Novel Multi-classification Approach for Drugs Based on ANN and Other ML algorithms with Hyper Parameter Tuning

A Novel Multi-Classification Approach for Drugs Based on ANN...

引用

Disruptive Technologies (ICDT), International Conference on

作者： Mohd Firoz Warsi Nitin Kumar Chauhan Shiv Narain Gupta Rahul Dev Ashish Choudhary Department of Computer Science (AIDS) Greater Noida Institute of Technology Greater Noida India Department of Elec. and Comm. Engg. Indore Institute of Science and Technology Indore India Department of Elec. and Comm. Engg. Greater Noida Institute of Technology Greater Noida India Department of Elec. and Comm. Engg. G.L. Bajaj Institute of Technology & Management Greater Noida Uttar Pradesh India

ISBN: (数字)9798331519582

ISBN: (纸本)9798331519599

In the realm of pharmaceutical research and development, accurate classification of drugs is of paramount importance for effective therapeutic interventions. This paper proposes a novel approach to drug classification utilizing a combination of Machine Learning (ML) and Artificial Neural Networks (ANN) algorithms. By harnessing the power of ANN's ability to learn complex patterns and ML's versatility in handling diverse datasets, our approach aims to enhance the accuracy and efficiency of drug classification tasks. Furthermore, we incorporate hyper parameter tuning techniques to optimize model performance, thereby maximizing predictive capabilities. Through comprehensive experimentation and evaluation on real-world drug datasets, our proposed methodology demonstrates promising results, showcasing superior classification accuracy compared to traditional methods. This research advances drug discovery and development by providing a robust framework for accurate drug classification in biomedical and pharmaceutical fields. In this study we have applied ten ML technique: Logistic Regression, Decision tree, Random forest, Catboost classifier, Adaboost classifier, Ensemble classifier, Gradient boosting classifier, LGBM classifier, Soft voting classifier, ANN with hyper parameter tuning and Support vector machine on the drugs classification dataset. The evaluation of classification outcomes was conducted both with and without the implementation of hyper parameter tuning. Notably, upon applying hyper parameter tuning, the ANN classifier emerged as the top performer, surpassing all other classifiers. RFC and DTC have also attained good accuracy of 96%. Impressively, ANN attained an accuracy rate of 99.29%, showcasing its superior performance in the dataset.

关键词： Drugs Support vector machines Logistic regression Accuracy Machine learning algorithms Artificial neural networks classification algorithms Drug discovery Tuning Random forests

来源：评论

学校读者我要写书评

暂无评论

Hybrid Optimization algorithms for Accurate Prediction and classification of Cardiovascular Diseases

Hybrid Optimization Algorithms for Accurate Prediction and C...

引用

Automation and Computation (AUTOCOM), International Conference on

作者： Gaurav Kumar Neeraj Varshney Department of Computer Engineering & Applications GLA University Mathura India

ISBN: (数字)9798331542375

ISBN: (纸本)9798331542382

A collection of precise prediction models is necessary to lessen the burden of cardiovascular disease (CVD), which continues to be a leading cause of death worldwide. This work presents a unique hybrid optimization approach for improving the categorization and prediction of cardiac disease. Support Vector Machine (SVM), Salp Swarm Optimized Neural Network (SSO-NN), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Bayesian Optimization of SVM (BO-SVM), and by combining PSO and gradient descent (PSOGD) are some of the advanced machine learning algorithms that have been implemented with improved accuracy. Training, testing, and validation are done using the Cleveland and Framingham datasets. Imputation for missing data and feature scaling using MinMaxScaler are important preprocessing processes. In order to evaluate performance parameters like as sensitivity, specificity, and accuracy. As a consequence, the hybrid PSOGD algorithm achieved impressive results, with accuracies with the dataset of Cleveland having 99.56% and other dataset Framingham having 98.80%. The model's robustness was shown by the sensitivity and specificity results. The measures like Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) over epochs, the performance analysis is further testified to for model validation. It suggests that hybrid optimization algorithms that improve CVD prediction models be praised as essential for management planning and early identification. By exposing scalable alternatives for heart disease prediction, our study offers crucial help in the area where AI and health intersect.

关键词： Heart Support vector machines Training Accuracy Sensitivity Computational modeling Predictive models classification algorithms Optimization Testing

来源：评论

学校读者我要写书评

暂无评论

Incorporating environmental data to refine the classification and understanding of the mechanisms behind encroachment of a woody species in the Southern Great Plains (USA)

引用

INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION 2025年 136卷

作者： Dawsey, Justin Mcintyre, Nancy E. Texas Tech Univ Dept Biol Sci 2901 Main St Lubbock TX 79409 USA

Curtailing encroachment is dependent on effectively identifying where problematic species occur. However, traditional classification methods struggle to distinguish spectrally similar species. New techniques that incorporate environmental variables (edaphic, climatic, and topographic characteristics) into classification can refine predictions and help identify important factors associated with species occurrence. We developed a workflow to improve classification of honey mesquite ( Neltuma [ = Prosopis ] glandulosa) in the Southern Great Plains (USA), examining 70 environmental variables to determine which were most associated with mesquite presence. We used Google Earth Engine to run X-means clustering on high-resolution aerial imagery from 50 replicate 78-km2 areas in New Mexico and Texas. We then refined our classification using XGBoost to generate accuracy assessment points for each area to confirm locations of mesquite clusters. Our method improved classification accuracy from 36 % to 83 %. We performed an ex-situ ground-truthed validation study and achieved 74 % accuracy. Inclusion of environmental data increased the accuracy of mesquite classification and allowed us to estimate the influence of each variable in determining whether a given point was classified as mesquite. Shallow, alkaline soils with low water-storage capacity, high electrical conductance, and low cation exchange capacity were associated with mesquite presence;these areas tended to be associated with flat, low-elevation drainages in regions that experience wide annual temperature ranges. These methods provide an easily reproducible and scalable way to assist with image classification of rangeland shrubs from remotely sensed imagery, which may prove useful in managing the further encroachment of problematic species like honey mesquite.

关键词： Tree-based models classification algorithms Conservation Woody plant encroachment

来源：评论

学校读者我要写书评

暂无评论

Reducing Data Volume in News Topic classification: Deep Learning Framework and Dataset

IEEE OPEN JOURNAL OF THE COMPUTER SOCIETY

引用

IEEE OPEN JOURNAL OF THE COMPUTER SOCIETY 2025年第1期6卷 152-163页

作者： Serreli, Luigi Marche, Claudio Nitti, Michele Univ Cagliari Dept Elect & Elect Engn DIEE I-09123 Cagliari Italy Natl Telecommun Inter Univ Consortium Res Unit Cagliari I-09123 Cagliari Italy

Withthe rise of smart devices and technological advancements, accessing vast amounts of information has become easier than ever before. However, sorting and categorising such an overwhelming volume of content has become increasingly challenging. This article introduces a new framework for classifying news articles based on a Bidirectional LSTM (BiLSTM) network and an attention mechanism. The article also presents a new dataset of 60 000 news articles from various global sources. Furthermore, it proposes a methodology for reducing data volume by extracting key sentences using an algorithm resulting in inference times that are, on average, 50% shorter than the original document without compromising the system's accuracy. Experimental evaluations demonstrate that our framework outperforms existing methodologies in terms of accuracy. Our system's accuracy has been compared with various works using two popular datasets, AG News and BBC News, and has achieved excellent results of 99.7% and 94.55%, respectively.

关键词： Feature extraction Text categorization classification algorithms Accuracy Bidirectional long short term memory Vectors Computational modeling Numerical models Nearest neighbor methods Radio frequency Data volume deep learning natural language processing topic classification

来源：评论

学校读者我要写书评

暂无评论

Flower classification: A Comparative Study of ML algorithms in Orange

Flower Classification: A Comparative Study of ML Algorithms ...

引用

Advanced Computing Technologies (ICoACT), International Conference on

作者： Rudraksh Agarwal Amritanshu Mukherjee Muskan Garg T R Saravanan N. Kanimozhi Department of Computational Intelligence S.R.M Institute of Science and Technology Kattankulathur Chennai

ISBN: (数字)9798331509859

ISBN: (纸本)9798331509866

This paper compares different machine learning models in Orange data mining tool to classify seven types of flowers using a dataset of about 140 images The study examines the performance of algorithms such as logistic regression, SVM, kNN, random forest, decision trees etc. based on key metrics such as accuracy, precision, recall, F1-score. The results indicate that Logistic Regression performed perfectly well 99.5%, followed by SVM with an accuracy of 98.9%. Random forest and kNN achieved lower accuracies of 89.9% and 78.7%, respectively. Decision trees had the lowest performance, with an accuracy of 54.7%. Logistic regression outperformed the other models in terms of accuracy and recall, thus becoming the most effective model for flower classification in this study. The results provide insights in choosing the best algorithms for this classification task.

关键词： Support vector machines Measurement Logistic regression Accuracy Flowering plants Data models classification algorithms Decision trees Data mining Random forests

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：