检索结果-内蒙古大学图书馆

Early detection of stroke disease using patients previous medical data instil with deep learning

Multimedia Tools and Applications 2025年第16期84卷 16853-16881页

作者： Diwan, Tausif Gajbhiye, Saurav M. Goydani, Purva R. Gannarpwar, Vedant R. Khandait, Harshal R. Tembhurne, Jitendra V. Sahare, Parul Department of Computer Science and Engineering Indian Institute of Information Technology Maharashtra Nagpur441108 India Department of Electronics and Communication Engineering Indian Institute of Information Technology Maharashtra Nagpur441108 India

Early detection of any disease and starting its treatment in this early stage are the most important steps in case of any life-threatening disease. Stroke is not an exception in this regard which is one of the leading causes of death and disability worldwide. We develop a simple but efficient deep neural network for the stroke prediction that accurately evaluates the probability of occurrence of stroke disease by treating this as a binary classification problem on one of the standard datasets named as Stroke Prediction Dataset available on Kaggle. With the help of effective pre-processing techniques such as SMOTE and FastICA on the noisy and imbalanced dataset, we could achieve improved performance for the binary classification of stroke prediction by employing a deep dense neural network. The architectural hyper-parameters are critically designed with the help of Keras Tuner and consisting of just ten layers in the deep dense neural network. With the help of this light weight deep dense model, we report an accuracy of 95%, AUC Score of 100% and F1-Score of 95% on a testing set. Moreover, we also present comparative illustration of various machine learning models for the aforesaid task along with the comparative illustration of various state of the arts with our proposed model for the stroke prediction. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Block Chain Based Voting System 2

Block Chain Based Voting System

引用

2nd IEEE International Conference for Women in Innovation, technology and Entrepreneurship, ICWITE 2024

作者： Patil, Vijeeta Bagalkot, Suhasini Rohit Guru, K.M. Kle Institute of Technology Computer Science and Engineering India Rrce Bangalore Computer Science and Engineering India

ISBN: (纸本)9798350383287

As we as a whole realize that casting ballot extortion is basic in India and furthermore in many created nations as well. There have been a few activities to lessen casting a ballot misrepresentation, for instance, the utilization of EVM machines in India. In the event that we can utilize some innovation to ensure against something similar. Here comes the utilization of Blockchain, a blockchain is a protected conveyed record that is constant and utilizes cryptographic techniques to execute the given properties. In this task, we plan a proposition for setting up the best answer for a blockchain casting a ballot decentralized applications(DApp) that consolidates citizen self-sovereign-ID and undeniable voting form arrangement. We will utilize brilliant agreements to execute the standards set up by the Election commission of India. Our proposition tends to the essential standards and shows key capacities and contemplations for a completely supported drive and improved elector end up, progressing commitment, and tremendous expense investment funds by limiting blunders and manual information passage and framework organization for heritage frameworks. We accept that an electronic democratic DApp ought to be reason-driven, the local area situated, in view of open guidelines for blockchain advancements, citizen protection, and security as a goal, and autonomous check and straightforwardness for polling from the organization. © 2024 IEEE.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Pseudo Random Number Generator Based on Cellular Automata with Self Organized Criticality

引用

SN computer science 2024年第5期5卷 454页

作者： Poornima, I. Gethzi Ahila Yogaraja, C.A. Venkatesh, R. Sudha, M. Swarna Vijayalakshmi, B. Computer Science and Engineering Ramco Institute of Technology Tamilnadu Rajapalayam India

In cryptography, pseudo-random numbers are crucial. The degree of strength of a cryptographic technique is directly influenced by the key’s randomness. Although numerous researchers have shown that cellular automata (CA) are effective as pseudo-random number generators (PRNGs), the fundamental CA structure is being changed to increase the efficiency in terms of randomness. Here, Langton's Ant Model and Programmable Controllable Cellular Automata (PCCA) in combination with a sand pile are proposed. Sand pile model and Langton’s Ant model are built for the generation of Rule control word for the successive iterations and Cell control word in selecting controllable cells for iterations respectively. The Controllable Cellular Automaton types used here are CCA4 and CCA5. Both Sand pile model and Langton’s Ant model exhibit chaotic behaviour. Two different rule sets are used and their results are compared. Also the frequency of each rule in the rule set are determined and found to be equally utilized. Due to the integration of these systems with cellular automata, which likewise display dynamic chaotic behaviour. Its results are greater efficiency in terms of randomness. Pseudo Random number sequence generated by this proposed PCCA is benchmarked with the statistical NIST test. The experiment shows that the resulting random sequence is highly random. Their entropies are calculated and the respective graphs are plotted. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： Cellular automata (CA) Cryptography Programmable Controllable Cellular Automata (PCCA) Pseudo Random number Pseudo-Random Number Generators (PRNGs)

来源：评论

学校读者我要写书评

暂无评论

mLBOA-DML: modified butterfly optimized deep metric learning for enhancing accuracy in intrusion detection system

引用

Journal of Reliable Intelligent Environments 2023年第3期9卷 333-347页

作者： Prabhakaran, Varun Kulandasamy, Ashokkumar Department of Computer Science and Engineering St.Joseph’s College of Engineering Chennai India Department of Computer Science and Engineering Sathyabama Institute of Science and Technology Chennai India

Intrusion detection is a prominent factor in the cybersecurity domain that prevents the network from malicious attacks. Cloud security is not satisfactory for securing the user’s information because it is based on standard protocols. Hence, cloud users cannot fully trust the security offered by cloud service providers. The state-of-the-art techniques create clusters for classes and manually label the unknown classes to detect novel attacks. This notion binds the network traffic associated with each attack together and drifts the similarity between the same attacks. These techniques are often prone to errors and degrade performance. To overcome this drawback, various researchers have developed different intrusion detection system which relies on specific attack patterns to distinguish between normal and abnormal behavior. This paper presents a modified Lagrange interpolated Butterfly optimization algorithm-based deep metric learning (mLBOA-DML) architecture for intrusion detection to detect both host-based and network attacks. DML architecture parameters are optimized utilizing mLBOA algorithm via its global optimization capability for increasing the accuracy of attack prediction. DML algorithm does both feature extraction and classification processes. When evaluated using the UNSW-NB15 and NSL-KDD datasets, the proposed model offers improved accuracy near 99% for both datasets. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.

关键词： Intrusion detection

来源：评论

学校读者我要写书评

暂无评论

A hybrid feature weighting and selection-based strategy to classify the high-dimensional and imbalanced medical data

引用

Neural Computing and Applications 2024年第20期36卷 12299-12316页

作者： Singh, Harpreet Kaur, Manpreet Singh, Birmohan Department of Computer Science and Engineering Sant Longowal Institute of Engineering and Technology Longowal Sangrur148106 India Department of Electrical and Instrumentation Engineering Sant Longowal Institute of Engineering and Technology Longowal Sangrur148106 India

Machine learning algorithms generally assume that the data are balanced in nature. However, medical datasets suffer from the curse of dimensionality and class imbalance problems. The medical datasets are obtained from the patient information which creates an imbalance in class distribution as the number of normal persons is more than the number of patients and contains a large number of features to represent a sample. It tends to the machine learning algorithms biased toward the majority class which degrades their classification performance for minority class samples and increases the computation overhead. Therefore, oversampling, feature selection and feature weighting-based four strategies are proposed to deal with the problems of class imbalance and high dimensionality. The key idea behind the proposed strategies is to generate a balanced sample space along with the optimal weighted feature space of the most relevant and discriminative features. The Synthetic Minority Oversampling Technique is utilized to generate the synthetic minority class samples and reduce the bias toward the majority class. An Improved Elephant Herding Optimization algorithm is applied to select the optimal features and weights for reducing the computation overhead and improving the interpretation ability of the learning algorithms by providing weights to relevant features. In addition, thirteen methods are developed from the proposed strategies to deal with the problems of high-dimensionality and imbalanced data. The optimized k-Nearest Neighbor (k-NN) learning algorithm is utilized to perform classification. The performance of the proposed methods is evaluated and compared for sixteen high-dimensional imbalanced medical datasets. Further, Freidman’s mean rank test is applied to show the statistical difference between the proposed methods. Experimental and statistical results show that the proposed Feature Weighting followed by the Feature Selection (FW–FS) method performed significantly b

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

Sem-AI: A Unique Framework for Sentiment Analysis and Opinion Mining Using Social Network Data

引用

SN computer science 2025年第2期6卷 1-23页

作者： Maruthupandi, J. Sivakumar, S. Kumar, V. Senthil Srikaanth, P. Balaji Department of Computer Science and Engineering New Horizon College of Engineering Karnataka Bangalore India Department of Computer Science and Engineering SRM Institute of Science and Technology Tamilnadu Trichy India Department of Networks and Communication SRM Institute of Science and Technology Tamilnadu Chennai India Department of Computer Science and Engineering Vel Tech Rangarajan Dr. Sagunthala R & D Institute of Science and Technology Chennai India

Social media has become an essential forum for people to share their thoughts and sentiments owing to the quick rise in mobile technology. Business and political organizations might benefit from understanding public sentiment when making strategic decisions. Challenges in the area include handling noisy and incomplete data, variation in language use across domains, and integration into existing systems of models developed for analysis. In these ways, they can seriously hamper the accuracy and generalizability of sentiment analysis models. In view of the challenges, this work has developed a new and innovative framework of AI-based semantic analysis for assessing and classifying user views from social data. In spite of this, sentiment analysis is crucial for determining the opposing perspectives of the general public's sentiments. This work constructed a new and innovative framework called Semantic analysis based on AI (SemAI) for assessing and categorizing user views from social data. With high efficiency, a collection of intelligence algorithms in the stages of feature extraction, optimization, and sentiment prediction are applied here. The Degree of Correlation Network Model (DCNM) is used in this work to extract the subset of dependable and most-required features from social network data. The most recent meta-heuristics model, Heap based Optimization (HbO), is used to reduce the dimensionality of retrieved features. Furthermore, the unique classification algorithm, Self-Attention based Deep Analyzing Network (SA-DAN), is used in this work to precisely identify and categorize attitudes into positive, neutral, and negative classifications. This present study used a variety of most recent social corpus data to investigate and evaluate the Sem-AI model's outputs. The SemAI model achieved impressive results, with accuracy, precision, and F1-scores of 99%, 98.9%, and 99%, respectively, across various social corpus datasets. © The Author(s), under exclusive licence to S

关键词： Classification Degree of correlation network model (DCNM) Heap based optimization (HbO) Opinion mining Self-attention based deep analyzing network (SA-DAN) Semantic analysis Social data

来源：评论

学校读者我要写书评

暂无评论

Prediction of banana leaf disease by intelligent algorithm through digital image using neural network technique

引用

Multimedia Tools and Applications 2025年 1-17页

作者： Nagu, Bharathiraja Kaur, Gaganpreet Shanmugavelu, Murugesan M.R.M, Veeramanickam Chitkara University Institute of Engineering and Technology Chitkara University Punjab India Department of Information Technology Tagore Engineering College Tamil Nadu Chennai India Apex Institute of Technology-Computer Science and Engineering Chandigarh University Punjab India

A lot of research shows that there could be several reasons why the duality of agricultural products has been reduced. Plant diseases make up one of the most important components of this quality. Therefore, the reduction of plant diseases allows for much higher product quality. The automated diagnosis of plant diseases using CNN techniques is presented in this article. The simple approach alters the 25 color channels and 24H histograms. The Banana Leaf Disease Detection Technique (BLDDT), which uses two attention mechanisms to assess, identify and categorize traits, was also suggested in this research. In the analysis phase, the testing carried out using the reference information reveals that the disease, detection system may be reasonably more appropriate than the current techniques because it both identifies and diagnoses the diseased/infected areas. The researchers were able to identify Esca, black decay, and isariopsis ––––with a 99.93% accuracy level using the suggested disease detector. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Integrating deep and handcrafted features for enhanced decision-making assistance in breast cancer diagnosis on ultrasound images

引用

Multimedia Tools and Applications 2025年 1-23页

作者： Abhisheka, Barsha Biswas, Saroj Kumar Purkayastha, Biswajit Das, Soumen School of Computer Science UPES Bidholi Uttrakhand Dehradun248007 India Computer Science Engineering National Institute of Technology Silchar NIT Road Assam Silchar788010 India

The classification of breast cancer has emerged as a significant concern in the healthcare sector in recent times. This is primarily due to its status as the second leading cause of cancer-related fatalities among women. Prior research in the field of breast cancer classification has mainly focused on either traditional Machine Learning (ML) based models or employing Deep Learning (DL) models. However, both approaches solely fail to capture the complete spectrum of features necessary for accurate classification. Therefore, this paper introduces a classification model named Hybrid Breast Cancer Prediction System (HBCPS), which utilizes a combination of features derived from deep Convolutional Neural Network (CNN) model and handcrafted features, to improve classification performance. The proposed HBCPS uses the pre-trained ResNet50 network to extract the deep features, while the handcrafted features are obtained using Histogram Orientation Gradient (HOG). For classification the proposed approach uses Support Vector Machine (SVM). In addition, the proposed method integrates the Block Matching 3D(BM3D) denoising filter. This filter efficiently reduces multiplicative noise such as speckle noise from Breast Ultrasound (BUS) images, resulting in improved image quality. This enhancement technique has significantly contributed to improve the overall performance of the system. The proposed framework is evaluated using a widely used Breast Ultrasound Image (BUSI) dataset. The proposed HBCPS exhibited satisfactory performance with an accuracy, precision, recall, F1-score, specificity, and AUC values of 89.02%, 87.67%, 87.17%, 87.36%, 83.87%, and 0.8717, respectively. These metrics indicate the robustness and reliability of the proposed method and offer a comprehensive framework for breast cancer classification using BUS images. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

The ringed residual u-net with non-natural regions feature for image splicing forgery detection and localization

引用

Journal of Intelligent and Fuzzy Systems 2024年第4期46卷 7447-7459页

作者： Wang, Qi Lu, Tongwei School of Computer Science and Engineering Wuhan Institute of Technology Wuhan China Hubei Key Laboratory of Intelligent Robot Wuhan Institute of Technology Wuhan China

Recently, with the emergence of many image editing tools (photoshop, Topaz studio, etc.), the authenticity of images has been severely challenged. However, the performance of some existing traditional feature extraction methods and detection methods based on convolutional neural network (CNN) is poor, and the information provided by the features extracted from the network is limited and single. In this paper, an end-To-end ringed residual U-Net is proposed to detect image splicing forgery by blending features of non-natural regions. Some regions with significant differences from the image background are defined as non-natural regions(such as the irregular border at the splicing of images). In this paper, a feature enhancement module for non-natural regions is constructed, which the image through the pooling of four different scales, and these features are then combined with the original image and input to the backbone network for processing, aiming to highlight regions of the image that differ significantly from the background. Therefore, after adding the feature enhancement module for non-natural regions to the end-To-end ring residual U-Net, more attention will be paid to the tampering regions in the feature extraction stage, image manipulation detection and localization will also become more accurate. Compared with some mainstream methods, this method achieves better performance on the three standard datasets(CASIA2.0, NIST2016, COLUMBIA). In addition, it has excellent robustness under JPEG compression attack and noise corruption attack. © 2024-IOS Press. All rights reserved.

关键词： Topaz

来源：评论

学校读者我要写书评

暂无评论

A topical VAEGAN-IHMM approach for automatic story segmentation

引用

Mathematical Biosciences and engineering 2024年第7期21卷 6608-6630页

作者： Yu, Jia Peng, Huiling Wang, Guoqiang Shi, Nianfeng School of Computer and Information Engineering Luoyang Institute of Science and Technology China Software Research Institute Technological University of Shannon Ireland

Feature representations with rich topic information can greatly improve the performance of story segmentation tasks. VAEGAN offers distinct advantages in feature learning by combining variational autoencoder (VAE) and generative adversarial network (GAN), which not only captures intricate data representations through VAE's probabilistic encoding and decoding mechanism but also enhances feature diversity and quality via GAN's adversarial training. To better learn topical domain representation, we used a topical classifier to supervise the training process of VAEGAN. Based on the learned feature, a segmentor splits the document into shorter ones with different topics. Hidden Markov model (HMM) is a popular approach for story segmentation, in which stories are viewed as instances of topics (hidden states). The number of states has to be set manually but it is often unknown in real scenarios. To solve this problem, we proposed an infinite HMM (IHMM) approach which utilized an HDP prior on transition matrices over countably infinite state spaces to automatically infer the state's number from the data. Given a running text, a Blocked Gibbis sampler labeled the states with topic classes. The position where the topic changes was a story boundary. Experimental results on the TDT2 corpus demonstrated that the proposed topical VAEGAN-IHMM approach was significantly better than the traditional HMM method in story segmentation tasks and achieved state-of-the-art performance. © 2024 the Author(s).

关键词： Hidden Markov models

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：