检索结果-内蒙古大学图书馆

International Conference on Electrical, computer and Communication engineering (ECCE)

作者： Ziyan Shirin Raha Md. Arafat Alam Khandaker Md. Zahid Hossain Shamim Akhter Department of Computer Science and Engineering Ahsanullah University of Science and Technology Dhaka Bangladesh

ISBN: (数字)9798350357509

ISBN: (纸本)9798350357516

This research presents a comprehensive analysis of various high-dimensional clustering algorithms applied to a large dataset comprising 5,000,000 data points. To manage the high dimensionality, Principal Component Analysis (PCA) reduces the datasets high dimensionality while keeping substantial variance. The study evaluates various clustering algorithms: DBSCAN, HDBSCAN, Spectral Clustering, BIRCH, Autoencoder + KMeans, KMedoids and CLARANS. Performance is assessed using Silhouette Score, Davies-Bouldin Score, and Calinski-Harabasz Index, measuring cluster compactness, separation and dispersion ratio, respectively. Our findings demonstrate that BIRCH excels with a high Silhouette Score of 0.9685, the lowest Davies-Bouldin Score of 0.0650, and a Calinski-Harabasz Index of 2105.4370, proving its ability in constructing well-defined, cohesive clusters. This study combines quantitative metrics and visual analyses to assist discover the best clustering approaches for high-dimensional data.

关键词： Measurement Visualization High dimensional data Navigation Autoencoders Clustering algorithms Intrusion detection Feature extraction Indexes Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Detection of Skin Cancer using CNN

Detection of Skin Cancer using CNN

引用

International Conference on Trends in Electronics and Informatics (ICEI)

作者： A. Ronald Doni Mahanthi Varun Mannepalli Manoj Kumar G. Nagarajan Department of computer science and engineering Sathyabama institute of science and technology Chennai India

ISBN: (数字)9798331544607

ISBN: (纸本)9798331544614

Skin diseases have been on the rise in human beings over the past few years. Ringworm, yeast infections, and allergies are typically caused by germs or infections. As infections can easily spread, it is imperative to detect them early to avoid further transmission. Clinical signs are necessary for the effective diagnosis of such illnesses. Various skin diseases affect individuals at different ages. For instance, children between the ages of 0 and 5 tend to have dermatitis, whereas those between the ages of 6 and 11 are most affected by acne. Teenagers between the ages of 12 and 16 are most susceptible to vulgaris. Children between the ages of 5 and 15 can also suffer from dermatomyositis. Adults aged between 40 and 60 also experience different skin diseases. Early detection and proper treatment are essential for tracking and containing the spread of these diseases.

关键词： Pediatrics Accuracy Image recognition Neural networks Market research Skin Convolutional neural networks Informatics Diseases Skin cancer

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Machine Intelligence Demographic Feature Selection Approach to Improve Recommendation System in Social Domain

A Hybrid Machine Intelligence Demographic Feature Selection ...

引用

International Conference on Machine Learning and Big Data Analytics, ICMLBDA 2023

作者： Vamsi, Bandi Mahanty, Mohan Sambana, Bosubabu Department of Artificial Intelligence Madanapalle Institute of Technology & Science Andhra Pradesh Madanapalle India Department of Computer Science and Engineering Vignan’s Institute of Information Technology Visakhapatnam India Makavarapalem Andhra Pradesh JNTU GV Vizianagaram India

ISBN: (纸本)9783031513374

Based on the preferences and previous history with many other customers, recommendation system assists users in quickly finding pertinent goods. They benefit users by making searching easier and companies by encouraging the sale of their products. Various filtering methods, such as collaborative, demographic, and content-based filtering are used to create a prediction model. Furthermore, because of concerns with variance and sparsity, massive amounts of information may result in suggestions with a limited degree of accuracy. In this work, we analyze random forest (RF) and AdaBoost classifiers and suggest a hybrid learning technique that uses user-to-user filtering features to identify the closest users. This technique is designed as a result of research into how to lower rating prediction mistakes based on user prior interactions, thus increasing prediction performance in two classification algorithms. We used a characteristic combination technique to increase predictability. Our proposed model has attained less error rate on using random forest and AdaBoost classifiers for both mean absolute error and root mean squared error filters. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

Explainable AI-Guided Neural Architecture Search for Adversarial Robustness in Approximate DNNs

IEEE Transactions on Sustainable Computing

引用

IEEE Transactions on Sustainable Computing 2025年

作者： Siddique, Ayesha Hoque, Khaza Anuarul University of Missouri Department of Electrical Engineering and Computer Science ColumbiaMO United States

Deep neural networks are lucrative targets of adversarial attacks and approximate deep neural networks (AxDNNs) are no exception. Searching manually for adversarially robust AxDNN architectures incurs outrageous time and human effort. In this paper, we propose XAI-NAS, an explainable neural architecture search (NAS) method that leverages explainable artificial intelligence (XAI) to efficiently co-optimize the adversarial robustness and hardware efficiency of AxDNN architectures on systolic-array hardware accelerators. During the NAS process, AxDNN architectures are evolved layer-wise with heterogeneous approximate multipliers to deliver the best trade-offs between adversarial robustness, energy consumption, latency, and memory footprint. The most suitable approximate multipliers are automatically selected from an open-source Evoapprox8b library. Our extensive evaluations provide a set of Pareto optimal hardware efficient and adversarially robust solutions. For example, a Pareto-optimal DNN AxDNN for the MNIST and CIFAR-10 datasets exhibits up to 1.5× higher adversarial robustness, 2.1× less energy consumption, 4.39× reduced latency, and 2.37× low memory footprint when compared to the state-of-the-art NAS approaches. © 2016 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Development of an Intrusion Detection Model Using a Support Vector Machine to Identify Cyber Attacks 4th

Development of an Intrusion Detection Model Using a Support ...

引用

4th International Research Conference on Machine Learning, Deep Learning, and Blockchain, IRCICD 2023

作者： Das, Pritam Singh, Akash Department of Computer Science and Engineering SRM Institute of Science and Technology Chennai India

ISBN: (纸本)9783031882364

Cyber-attacks are deliberate attempts to gain unauthorised access to a person’s computer and data. As the web has grown, it has become a fundamental part of most people’s daily routine, and there are concerns about online security. In addition, the growing amount of information being shared between clouds and clients creates an attack surface that can be exploited. In addition to the increase in organizations and users, the attack surface has also increased. Deterioration in existing discovery plans is resulting in a reluctance to approve goals and an earlier acknowledgment of assaults. If no efficient assurance system is put in place, the web will grow increasingly susceptible, increasing the risk of information leakage. This research suggests an Intrusion Detection System (IDS) that divides packet traffic into attack and non-attack classes and uses Support Vector Machines to categorise malicious classes to identify network intrusions. The Support Vector Machine (SVM) is considered one of the most effective and promising techniques in machine learning Algorithms. On decision boundaries, Support Vector Machine Classification (SVMC) is founded. It supports all binary and multi-class classifications. KDDCUP 1999 and UNSW NB15 are two different cyber security datasets that we utilised in our research. Performance criteria have been used to assess the suggested model. The test results show that our methodology is more effective at identifying different cyberattacks. For the UNSW NB15 dataset and KDDCUP 1999 dataset, this model obtains remarkable attack detection rates of 99.8% and 98.2%, respectively. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

The Impact of Phase Information on the Imaging Performance of the Extended Phaseless Rytov Approximation

IEEE Open Journal of Antennas and Propagation

引用

IEEE Open Journal of Antennas and Propagation 2025年

作者： Li, Zan Dubdy, Amartansh Murch, Ross Hong Kong University of Science and Technology Department of Electronic and Computer Engineering Hong Kong

The extended phaseless Rytov approximation for lossy media (xPRA-LM) has been shown to significantly extend the validity range of the Rytov approximation. Due to its extended validity range, xPRA-LM can be useful for large-scale RF imaging applications such as Wi-Fi based indoor imaging. However xPRA-LM is formulated for datasets with phaseless data only. In this work, we investigate the extended phaseless Rytov approximation when phase is included and we refer to this formulation as the extended full-data Rytov approximation for low-loss media (xFRA-LM). Specific contributions of this work include the formulation of xFRA-LM, development of a Cramér–Rao Bound for xFRA-LM and also providing numerical comparisons with xPRA-LM. The development of xFDA-LM also opens a pathway for the extension of radio tomographic imaging (RTI) to include phase and this is also discussed. The results show that the Bayesian Cramér–Rao Bound for xFRA-LM is at least half that of the bound for xPRA-LM demonstrating the advantages of including phase. Furthermore, simulation results for reconstruction quality show that xFRA-LM is significantly enhanced as compared to xPRA-LM, in terms of validity range, imaging accuracy and imaging resolution. © 2020 IEEE.

关键词： Cramer-Rao bounds

来源：评论

学校读者我要写书评

暂无评论

Water Distribution Network for Irrigation using Artificial Neural Network (ANN) Algorithm

Water Distribution Network for Irrigation using Artificial N...

引用

2025 International Conference on Visual Analytics and Data Visualization, ICVADV 2025

作者： Sankari, M. Viveha, R. Varshita, Y. Sathyabama Institute of Science and Technology Department of Computer Science and Engineering Chennai India

ISBN: (数字)9798331521394

ISBN: (纸本)9798331521394

With agriculture being a significant consumer of freshwater resources, optimizing water distribution based on accurate soil moisture predictions can minimize water waste, preserving valuable resources for other needs and reducing the environmental footprint of farming. Proper irrigation management based on real-time soil moisture predictions ensures crops receive adequate water at critical growth stages, improving both yield and quality, which is essential for food security and economic viability. This project presents the development of an Artificial Neural Network (ANN) model for optimizing a water distribution network for irrigation. The ANN model predicts soil moisture levels based on geospatial data, soil properties, and temporal features, enabling efficient irrigation scheduling. The dataset used includes soil texture properties (clay, sand, and silt content), geospatial coordinates (latitude and longitude), auxiliary soil moisture measurements, and temporal features (month and day of the year). The model architecture includes multiple hidden layers with ReLU activation functions and a linear output layer to predict the target soil moisture content. The system aims to optimize water usage by accurately predicting soil moisture, contributing to more sustainable irrigation practices. © 2025 IEEE.

关键词： Soil moisture

来源：评论

学校读者我要写书评

暂无评论

ADVANCING RETRIEVAL-AUGMENTED GENERATION FOR PERSIAN: DEVELOPMENT OF LANGUAGE MODELS, COMPREHENSIVE BENCHMARKS, AND BEST PRACTICES FOR OPTIMIZATION

arXiv

引用

arXiv 2025年

作者： Hosseinbeigi, Sara Bourbour Asghari, Sina Kashani, Mohammad Ali Seif Shalchian, Mohammad Hossein Abbasi, Mohammad Amin Department of Industrial and Systems Engineering Tarbiat Modares University Tehran Iran Department of Computer Science Iran University of Science and Technology Tehran Iran Department of Computer Engineering Sharif University of Technology Tehran Iran Department of Computer Engineering Iran University of Science and Technology Tehran Iran

This paper examines the specific obstacles of constructing Retrieval-Augmented Generation (RAG) systems in low-resource languages, with a focus on Persian’s complicated morphology and versatile syntax. The research aims to improve retrieval and generation accuracy by introducing Persian-specific models, namely MatinaRoberta (a masked language model) and MatinaSRoberta (a fine-tuned Sentence-BERT), along with a comprehensive benchmarking framework. Three datasets—general knowledge (PQuad), scientifically specialized texts, and organizational reports—were used to assess these models after they were trained on a varied corpus of 73.11 billion Persian tokens. The methodology involved extensive pretraining, fine-tuning with tailored loss functions, and systematic evaluations using both traditional metrics and the Retrieval-Augmented Generation Assessment (RAGAS) framework. The results show that MatinaSRoberta outperformed previous embeddings, achieving superior contextual relevance and retrieval accuracy across datasets. Temperature tweaking, chunk size modifications, and document summary indexing were explored to enhance RAG setups. Larger models like Llama-3.1 (70B) consistently demonstrated the highest generation accuracy, while smaller models faced challenges with domain-specific and formal contexts. The findings underscore the potential for developing RAG systems in Persian through customized embeddings and retrieval-generation settings and highlight the enhancement of NLP applications such as search engines and legal document analysis in low-resource languages. © 2025, CC BY-NC-SA.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Bangla Book Genre Classification: An Advanced Multi-Class Approach Using Transformer-Based Models

Bangla Book Genre Classification: An Advanced Multi-Class Ap...

引用

International Conference on Electrical, computer and Communication engineering (ECCE)

作者： Md. Shohanuzzaman Md. Nihan Ali Mohammad Shahidur Rahman Department of Computer Science and Engineering Shahjalal University of Science and Technology Sylhet Bangladesh

ISBN: (数字)9798350357509

ISBN: (纸本)9798350357516

This paper explores a diverse range of multiclass classification methods for Bangla books, particularly those accessible on digital platforms. Despite being one of the most widely spoken languages with a rich literary heritage, Bangla remains underexplored in content classification due to the lack of a comprehensive dataset. We introduce an expanded dataset of 46,429 books classified into 14 distinct categories. The categories are Education, Fiction, Non-Fiction, Poetry, Children’s Book, History, Biography, Literature, Adventure, Crime, Sci-Fi, Programming, Others. In order to achieve this, we used advanced machine learning models like BanglaBERT, XLM-RoBERTa and a CNN-LSTM attention-based deep learning approach. Our models demonstrate substantial improvements in classification accuracy, attributed to the increased dataset size and the adoption of refined model architectures. Both attention base and transformer based have outperformed previous benchmarks, with accuracy leaping from approximately around 16% to around 75% a direct result of the novel dataset we *** the models that we have tested, BanglaBERT stands out in terms of stability, showcasing exceptional performance due to its advanced transformer-based architecture. This positions BanglaBERT as the most effective model for Bangla content classification in our study, highlighting its superiority in handling the linguistic complexities of Bangla text.

关键词： Accuracy Computational modeling Biological system modeling Text categorization computer architecture Linguistics Transformers Stability analysis Complexity theory Thermal stability

来源：评论

学校读者我要写书评

暂无评论

Revolutionizing Cardiovascular Disease Prediction: Stacked Ensemble Models with XGBoost, Neural Networks, and Generative AI Integration

Revolutionizing Cardiovascular Disease Prediction: Stacked E...

引用

Data science, Agents & Artificial Intelligence (ICDSAAI), International Conference on

作者： N R Sathis Kumar K Baalaji Sivagnanasekar G Vijayarajan S Shameem Mohamed A R Hari Madhavan C Department of Computer Science and Engineering Kalasalingam Academy of Research & Education Krishnankovil TamilNadu Department of Computer Science and Engineering Bharath Institute of Science and Technology Bharath Institute of Higher Education & Research

ISBN: (数字)9798331537555

ISBN: (纸本)9798331537562

cardiovascular disease hence makes a sizable portion of global mortality, further calling for ideal and timely prediction measures. The comparison of four machine learning models, namely, Logistic Regression, Random Forest, XGBoost, and Neural Networks for cardiovascular disease prediction is done in this work. Therefore, to improve the prediction accuracy still further, the present study advocated the use of a new Stacking Model as the augmented learner through Random Forest and XGBoost algorithms. When compared with individual models, the accuracy of the Stacking Model is extraordinary is 90.00% higher than other baselines including Logistic Regression (72.36%), Random Forest (71.41%), XGBoost (73.57%), and Neural Networks (73.18). Within the system setting, there is a Generative AI chatbot where users can ask questions about the predictions and factors behind them. Moreover, a user-oriented feature implemented via an SMTP based Email Notification Service means timely alerting of severe predictions and/or changes. The combination of machine learning, Generative AI, and email notification shows how more prompt and efficient, easy to use diagnostic tools can be created for the future with appeal to personalized health solutions.

关键词： Logistic regression Accuracy Generative AI Neural networks Stacking Medical services Predictive models Chatbots Cardiovascular diseases 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：