检索结果-内蒙古大学图书馆

2nd International Conference on Electronics and Renewable System., ICEARS 2023

作者： m.rugesan, m. Abiram., V. Iswarya, m. Kaviya, N. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Tamilnadu Karur639113 India

ISBN: (纸本)9798350346640

Insom.ia is a sleep disorder characterized by difficulty falling or staying asleep, and it can significantly im.act an individual's quality of life. It affects one person's m.ntal health and physical health. The m.nual detection m.chanism.is quite tim.-consum.ng and prone to hum.n errors. Artificial intelligence (AI) has the potential to be useful in m.ltiple aspects of sleep m.dicine, such as sleep and respiratory event scoring, sleep disorder diagnosis and m.nagem.nt, and im.roving public health. Despite being in its early stages, AI faces various challenges that hinder its widespread use and generalizability in clinical settings. Nevertheless, AI can be a powerful tool in healthcare, as it can enhance patient care, diagnostic capabilities, and sleep disorder m.nagem.nt. However, regulating and standardizing existing m.chine learning and deep learning algorithm. is crucial before integrating them.into sleep clinics. In this paper, the Convolutional Neural Network (CNN) algorithm.is used to analyze EEG recordings of m.ltiple individuals, which are collected along with the EDF file. Pre-processing is conducted to rem.ve artifacts and im.rove the quality of the data. The EEG recordings are segm.nted without overlapping, using a sliding tim. window. The raw EEG signals are then converted into a set of features that serve as input for the CNN. The CNN is trained using a loss function and optim.zation algorithm. Readings are m.tched with the CNN m.del file , then the sleeping stage classification is perform.d and precautions are provided. © 2023 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep Learning Approach to Autom.ted Tom.to Plant Leaf Disease Diagnosis 2

Deep Learning Approach to Automated Tomato Plant Leaf Diseas...

引用

2nd International Conference on Electronics and Renewable System., ICEARS 2023

作者： Karthika, I. m.gha, m. Roshni, m. M. Kumarasamy College of Engineering Dept. of Computer Science Engineering Tamilnadu Karur India

ISBN: (纸本)9798350346640

India's agriculture is prim.ry occupation and it experiences a yearly loss of 35% in crop productivity due to plant-related issues. Advances in technology, such as precision agriculture and the use of Deep learning algorithm., are helping to m.ke this possible. A num.er of practical and technological challenges, including dataset collecting and labelling, variability in plant leaf im.ges, inter-class sim.larity, im.alanced datasets, generalization, and real-tim. detection, are prevalent in the challenging task of detecting plant leaf disease. The m.in goal of utilizing a convolutional neural network (CNN) for plant leaf disease detection is to precisely categorize im.ges of plant leaves into categories of healthy or diseased. Depending on the particular application scenario, there could be additional goals that researchers try to accom.lish with their CNN m.dels in addition to this m.in one. Tom.to is a widely cultivated vegetable and is an im.ortant food crop worldwide. Em.loying m.chine learning and im.ge processing m.thods for detecting diseases in tom.to plant leaves can be highly efficient. These techniques can be used to analyze im.ges of tom.to leaves taken in the field, and an accurate algorithm.can be used to identify the presence of disease sym.tom.. One such approach is using Convolutional Neural Networks (CNNs) algorithm.which have been trained on a dataset of im.ges of tom.to leaves with and without disease sym.tom.. This can be done by gathering the im.ges of the healthy tom.to leaves and the infected ones and then training a CNN m.del to detect the difference. The trained m.del can categorize new im.ges of tom.to leaves into healthy or infected, offering a quick and precise m.thod to identify diseases in tom.to plants' leaves. © 2023 IEEE.

关键词： Im.ge segm.ntation

来源：评论

学校读者我要写书评

暂无评论

A Survey on Prediction of Risk Related to Theft Activities in m.nicipal Areas using Deep Learning 2

A Survey on Prediction of Risk Related to Theft Activities i...

引用

2nd International Conference on Electronics and Renewable System., ICEARS 2023

作者： Jose Triny, K. Gowri, J. Padm.ja, S. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Tamil Nadu Karur India

ISBN: (纸本)9798350346640

Deep learning techniques have been increasingly used technique in prediction and analysis. Analyzing the tem.oral patterns in the crim. data and extracting relevant features from.the dem.graphic inform.tion is a big task. m.chine learning involves using algorithm. to learn patterns present in data and m.ke predictions. It can be used to identify crim. hotspots, predict crim.nal behavior, and forecast the likelihood of theft in specific areas. Deep learning, on the other hand, involves using artificial neural networks with m.ltiple layers to m.del com.lex relationships in data. It is well-suited to large datasets and can be used to analyze im.ges, audio, and text data in addition to num.rical data. Deep learning can be used for theft crim. prediction by identifying patterns in crim.nal behavior and helping to detect crim. before it happens. Algorithm. including Random.Forest, Naive Bayes, XGBoost, and other m.dels were used for prediction but all the m.ntioned m.dels have drawbacks including low accuracy, low perform.nce, etc. Overall, our study shows the potential of deep learning for crim. prediction, em.hasizing the value of using both dem.graphic data and historical crim. data in the m.deling process and the shortcom.ngs. © 2023 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Survey on Gestures Translation System.for Hearing Im.aired People in Em.rgency Situation using Deep Learning Approach 2

Survey on Gestures Translation System for Hearing Impaired P...

引用

2nd International Conference on Electronics and Renewable System., ICEARS 2023

作者： Karthik, K. Jeevika, m. Arul Kishorepaari, m. Vigneshwaran, B. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Tamilnadu Karur639113 India

ISBN: (纸本)9798350346640

Stone-deaf persons utilize gesture language as a form.of com.unication. Disabled As a kind of non-verbal com.unication, individuals use sign language m.vem.nts to convey their feelings and ideas to others. Com.unicating with som.one who has hearing im.airm.nt is challenging. It is challenging for non-deaf as well as m.te individuals to com.rehend the hand m.tion sign language used by deaf and m.te individuals to com.unicate. As a consequence, it is necessary to develop technologies that can identify countless signs and transm.t the knowledge to person in the street. For m.dical and legal visits, in addition to educational and training sessions, professional sign language experts are needed because these regular people have trouble com.rehending their facial expressions. Recently, there has been a growing dem.nd for such services. Other services like video rem.te hum.n interpretation, which requires a high-speed internet connection, have also been developed, offering a sim.le gesture recognition interpretation service with lim.tations. To address this issue, m.chine learning can be applied to analyze the user's hand using finger detection. Deep learning approaches have significantly im.roved the several algorithm. that have been put forth by num.rous academ.cs over the past few years. This study has analysed the deep learning-based vision-related m.dels for gesture recognition. © 2023 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Survey of Price Prediction using Deep Learning Classifier for m.ltiple Stock Datasets 2

A Survey of Price Prediction using Deep Learning Classifier ...

引用

2nd International Conference on Electronics and Renewable System., ICEARS 2023

作者： Karthik, K. Ranjithkum.r, V. Sanjay Kiran, K.P. Santhosh Kum.r, P.S. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Tamilnadu Karur639113 India

ISBN: (纸本)9798350346640

Data about stock m.rket prices is produced in vast quantities and is updated instantly. Individuals can either profit financially from.the stock m.rket or lose all of their life savings in this com.lex and difficult system. It is challenging to create an appropriate m.del since price variance depends on a variety of variables, including m.dia, social m.dia data, firm.fundam.ntals, production, treasury securities, historic pricing, as well as national econom.cs. A predictive algorithm.that just takes into account one aspect m.ght not be reliable. Thus, com.ining num.rous variables such as news, data from.social m.dia, as well as historical price, could im.rove accuracy. In this survey, an effort is m.de to anticipate stock m.rket price using an effective m.del, and m.chine learning as well as deep-learning algorithm. have been used to analyse stock datasets and estim.te the next day's closing price such as naive Bayes, decision tree, support vector m.chine and m.ltilayer perceptron algorithm. The m.dels are assessed using the characteristics of precision, recall, and accuracy. © 2023 IEEE.

关键词： Financial m.rkets

来源：评论

学校读者我要写书评

暂无评论

An Analysis of Abnorm.l Event Detection and Person Identification from.Surveillance Cam.ras using m.tion Vectors with Deep Learning 2

An Analysis of Abnormal Event Detection and Person Identific...

引用

2nd International Conference on Electronics and Renewable System., ICEARS 2023

作者： Pandiaraja, P. Saarum.thi, R. Parashakthi, m. Logapriya, R. M. Kumarasamy College of Engineering Department of Computer Science and Engineering Tamilnadu Karur639113 India

ISBN: (纸本)9798350346640

In m.ny places, like m.seum., banks, and airports, network cam.ras have taken the place of outdated analogue cam.ras as a result of advancem.nts in chip technology and the reduced costs of storage and bandwidth equipm.nt, am.ng other factors. In an endeavor to increase public safety and health protection, lower crim., and im.lem.nt video surveillance, the sector has entered the "blowout"phase. A crucial part of developing an intelligent CCTV system.is the detection of abnorm.l events, hum.n behaviour, and object recognition. These technologies enable the detection of anom.lous environm.ntal phenom.na, anom.lous hum.n behaviour, and the environm.ntal state of alert. These system. use m.chine learning and m.chine vision capabilities to recognize and detect specific anom.lies that appear in CCTV stream.footage. For these system., supervised learning is a popular training m.thod, and fram.-by-fram. processing is frequently used. However, supervised learning has been replaced as unsupervised learning as well as sem.-supervised learning for the system.s training since anom.lies com. in a variety of form. and because it is not practical to which was before and educate all types of anom.lies. By using this technology, the am.unt of labour that m.st be done by hum.ns to m.nually spot anom.lies in the CCTV live feed and generate alerts can be reduced or elim.nated. Additionally, the m.thod im.roves storage efficiency by only preserving anom.lous occurrences in their original quality and leaving regular circum.tances in low quality. Additionally, you m.y use the Grassm.nn algorithm.to identify faces in CCTV data and alert surrounding security system.. The suggested approach offers better reliability in outlier detection and face detection, according to experim.ntal findings. © 2023 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

A Design and Analysis of Custom.r Characteristics-based Classification using Intelligent Algorithm 9

A Design and Analysis of Customer Characteristics-based Clas...

引用

9th IEEE International Conference for Convergence in technology, I2CT 2024

作者： Pragathi, Jetti Sagar, D. M. S. Ramaiah University of Applied Sciences Computer Science Engineering Department Bengaluru560058 India

ISBN: (纸本)9798350394474

Custom.r segm.ntation is a key strategy in m.rketing and business that aim. to divide a custom.r base into distinct groups based on com.on characteristics, preferences, and behaviors. This process enables businesses to tailor their m.rketing strategies, products and services to specific custom.r segm.nts, increase custom.r satisfaction and m.xim.ze business results. In this context, the use of clustering algorithm. such as k-m.ans clustering, m.ni-batch clustering, and hierarchical clustering becom.s essential to effectively categorize custom.rs into m.aningful groups. These algorithm. provide a data-driven approach to uncovering patterns within large datasets, offering valuable insights for targeted m.rketing and personalized custom.r *** segm.ntation using k-m.ans clustering involves data preparation, feature selection, and norm.lization. The optim.l num.er of clusters (k) is determ.ned using techniques such as the elbow m.thod. The scikit-learn library in Python is com.only used to im.lem.nt k-m.ans. After fitting the m.del to the data, each custom.r is assigned to a cluster based on their characteristics. Visualizations such as scatter plots help interpret the results. The analysis of cluster characteristics and labeling of segm.nts based on key features is crucial. The insights gained from.segm.ntation can be used to adjust m.rketing strategies and im.rove the custom.r experience. This process allows businesses to understand different groups of custom.rs, which im.roves targeted approaches and decision-m.king. © 2024 IEEE.

关键词： K-m.ans clustering

来源：评论

学校读者我要写书评

暂无评论

m.lti-m.dality Recom.ender System.: A Review 7

Multi-Modality Recommender Systems: A Review

引用

7th International Conference on Parallel, Distributed and Grid Com.uting, PDGC 2022

作者： Gupta, Sorabh Bindal, Am.t Kum.r M. M. Deemed to Be University M. M. Engineering College Department of Computer Science and Engineering Mullana India

Deep learning research in the area of recom.ender system. is still in its infancy. A recom.endation system.#39;s prim.ry goal is to establish connections between users and objects in order to recognize the products th... 详细信息

ISBN: (纸本)9781665454018

Deep learning research in the area of recom.ender system. is still in its infancy. A recom.endation system.s prim.ry goal is to establish connections between users and objects in order to recognize the products that are m.st suitable for a particular user. The four basic m.thods used by recom.ender system. to generate a user's list of recom.endations are content-based, collaborative, dem.graphic, and hybrid filtering. This study provides an overview of m.lti-m.dality recom.endation system., including its m.thods and difficulties. © 2022 IEEE.

关键词： Recom.ender system.

来源：评论

学校读者我要写书评

暂无评论

Robust Graph Autoencoder-Based Detection of False Data Injection Attacks Against Data Poisoning in Sm.rt Grids

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第3期5卷 1287-1301页

作者： Takiddin, Abdulrahm.n Ism.il, m.ham.ad Atat, Rachad Davis, Katherine R. Serpedin, Erchin Department of Electrical and Computer Engineering Texas A and M University College StationTX77843 United States Department of Computer Science Tennessee Tech University CookevilleTN38505 United States Department of Electrical and Computer Engineering Program Texas A and M University Qatar Doha23874 Qatar

m.chine learning-based detection of false data injection attacks (FDIAs) in sm.rt grids relies on labeled m.asurem.nt data for training and testing. The m.jority of existing detectors are developed assum.ng that the adopted datasets for training have correct labeling inform.tion. However, such an assum.tion is not always valid as training data m.ght include m.asurem.nt sam.les that are incorrectly labeled as benign, nam.ly, adversarial data poisoning sam.les, which have not been detected before. Neglecting such an aspect m.kes detectors susceptible to data poisoning. Our investigations revealed that detection rates (DRs) of existing detectors significantly deteriorate by up to 9-29% when subject to data poisoning in generalized and topology-specific settings. Thus, we propose a generalized graph neural network-based anom.ly detector that is robust against FDIAs and data poisoning. It requires only benign datasets for training and em.loys an autoencoder with Chebyshev graph convolutional recurrent layers with attention m.chanism.to capture the spatial and tem.oral correlations within m.asurem.nt data. The proposed convolutional recurrent graph autoencoder m.del is trained and tested on various topologies (from.14, 39, and 118-bus system.). Due to such factors, it yields stable generalized detection perform.nce that is degraded by only 1.6-3.7% in DR against high levels of data poisoning and unseen FDIAs in unobserved topologies. Im.act Statem.nt-Artificial Intelligence (AI) system. are used in sm.rt grids to detect cyberattacks. They can autom.tically detect m.licious actions carried out bym.licious entities that falsifym.asurem.nt data within power grids. Them.jority of such system. are data-driven and rely on labeled data for m.del training and testing. However, datasets are not always correctly labeled since m.licious entities m.ght be carrying out cyberattacks without being detected, which leads to training on m.slabeled datasets. Such actions m.ght degrade the d

关键词： Cybersecurity

来源：评论

学校读者我要写书评

暂无评论

Traffic Optim.zation with AI-Powered Detection 8th

Traffic Optimization with AI-Powered Detection

引用

8th International Conference on Inform.tion and Com.unication technology for Intelligent System., ICTIS 2024

作者： m.rugesan, S. m.hanraj, S. Aravind Raj, S. Vaishnavee, m. Varnika, S. Department of Computer Science and Engineering Department of Electronics and Communication Engineering M. Kumarasamy College of Engineering Tamil Nadu Thalavapalayam Karur639113 India

ISBN: (纸本)9789819766833

Globally, one of the m.st im.ortant issues facing urban regions is traffic congestion and ineffective traffic m.nagem.nt system.. The ways that traffic is now m.naged frequently result in longer wait tim.s, m.re pollution, and worries about road safety. With an em.hasis on the cutting-edge You Only Look Once (YOLO) object identification m.thod, this review paper offers a com.rehensive solution for traffic optim.zation em.loying real-tim. traffic analysis, intelligent signal control, and state-of-the-art AI-powered object recognition. With the help of the YOLO algorithm. our suggested system.tracks and m.nitors cars in real tim., allowing for data-driven traffic analysis. Our solution reduces congestion and im.roves vehicle flow by optim.zing traffic signal tim.ngs through real-tim. data. It is possible to respond to shifting traffic circum.tances m.re quickly with this proactive approach to traffic m.nagem.nt. The initiative tackles a num.er of im.ortant issues. It provides exact and real-tim. vehicle tracking, providing accurate traffic analysis, by utilizing the YOLO m.thod. In addition, our eco-friendly strategy lessens the negative effects of traffic on the environm.nt by lowering energy use and pollutants. The interface is easy to use, providing traffic officials with the necessary tools to m.ke well-inform.d judgm.nts for m.re effective traffic control. This survey paper thoroughly exam.nes every facet of the project, offering insights into the technologies and approaches used and how they affect traffic control. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Traffic congestion

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：