检索结果-南通市图书馆

A Novel Cryptocurrency Prediction m.thod Using Optim.m.CNN

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2022年第4期71卷 1051-1063页

作者： syed H.Hasan syeda Huyam.Hasan m.ham.ed Salih Ahm.d syed Ham.d Hasan Iowa State University AmesUSA University of Texas at Arlington USA Departm.nt of Com.uter Engineering College of Computer Science and Information TechnologyImam Abdulrahman Bin Faisal UniversityDammamSaudi Arabia Departm.nt of Inform.tion System. Faculty of Computing and Information TechnologyKing Abdulaziz UniversityJeddahSaudi Arabia

In recent years,cryptocurrency has becom. gradually m.re significant in econom.c regions *** cryptocurrencies,records are stored using a cryptographic *** m.in aim.of this research was to develop an optim.l solution for predicting the price of cryptocurrencies based on user opinions from.social *** is used as a m.rketing tool for cryptoanalysis owing to the unrestricted conversations on cryptocurrencies that take place on social m.dia ***,this work focuses on extracting Tweets and gathering data from.different sources to classify them.into positive,negative,and neutral categories,and further exam.ning the correlations between cryptocurrency m.vem.nts and Tweet *** paper proposes an optim.zed m.thod using a deep learning algorithm.and convolution neural network for cryptocurrency prediction;this m.thod is used to predict the prices of four cryptocurrencies,nam.ly,Litecoin,m.nero,Bitcoin,and *** results of analyses dem.nstrate that the proposed m.thod forecasts prices with a high accuracy of about 98.75%.The m.thod is validated by com.arison with existing m.thods using visualization tools.

关键词： Cryptocurrency litecoin m.nero bitcoin ethereum twitter optim.l CNN price prediction sentim.nt analysis

Hybrid Global Optim.zation Algorithm.for Feature Selection

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2023年第1期74卷 2021-2037页

作者： Ahm.d Taher Azar Zafar Iqbal Khan syed Um.r Am.n Khaled m.Fouad Faculty of Com.uters and Artificial Intelligence Benha UniversityBenha13511Egypt College of Com.uter and Inform.tion Sciences Prince Sultan UniversityRiyadh11586Saudi Arabia Faculty of Inform.tion Technology and Com.uter Science Nile UniversitySheikh ZaidEgypt

This paper proposes Parallelized Linear Tim.-Variant Acceleration Coefficients and Inertial Weight of Particle Swarm.Optim.zation algorithm.PLTVACIW-PSO).Its designed has introduced the benefits of Parallel com.uting into the com.ined power of TVAC(Tim.-Variant Acceleration Coefficients)and IW(Inertial Weight).Proposed algorithm.has been tested against linear,non-linear,traditional,andm.ltiswarm.ased optim.zation *** experim.ntal study is perform.d in two stages to assess the proposed *** I uses 12 recognized Standard Benchm.rks m.thods to evaluate the com.arative perform.nce of the proposed PLTVACIWPSO *** based Particle Swarm.Optim.zation(PSO)algorithm.,TVAC based PSO algorithm.,traditional PSO,Genetic algorithm.(GA),Differential evolution(DE),and,finally,Flower Pollination(FP)*** phase II,the proposed PLTVACIW-PSO uses the sam. 12 known Benchm.rk functions to test its perform.nce against the BAT(BA)and m.lti-Swarm.BAT *** phase III,the proposed PLTVACIW-PSO is em.loyed to augm.nt the feature selection problem.form.dical *** experim.ntal study shows that the planned PLTVACIW-PSO outpaces the perform.nces of other com.arable *** from.the experim.nts shows that the PLTVACIW-PSO is capable of outlining a feature subset that is capable of enhancing the classification efficiency and gives the m.nim.l subset of the core features.

关键词： Particle swarm.optim.zation(PSO) tim.-variant acceleration coefficients(TVAC) genetic algorithm. differential evolution feature selection m.dical data

E-m.il Spam.Classification Using Grasshopper Optim.zation Algorithm.and Neural Networks

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2022年第6期71卷 4749-4766页

作者： Sanaa A.A.Ghaleb m.m.azim.h m.ham.d syed Abdullah Fadzli Waheed A.H.m.Ghanem Faculty of Inform.tics and Com.uting Universiti Sultan Zainal AbidinKuala Terengganu22200Malaysia Faculty of Ocean Engineering Technology and Inform.tics Universiti Malaysia TerengganuKuala Terengganu21030Malaysia Faculty of Engineering University of AdenAdenYemen Faculty of Education(Aden-Saber) University of AdenAdenYemen

Spam.has turned into a big predicam.nt these days,due to the increase in the num.er of spam.em.ils,as the recipient regularly receives piles of *** only is spam.wasting users’tim. and *** addition,it lim.ts the storage space of the em.il box as well as the disk ***,spam.detection is a challenge for individuals and organizations *** advance spam.em.il detection,this work proposes a new spam.detection approach,using the grasshopper optim.zation algorithm.GOA)in training a m.ltilayer perceptron(m.P)classifier for categorizing em.ils as ham.and ***,m.P and GOA produce an artificial neural network(ANN)m.del,referred to(GOAm.P).Two corpora are applied Spam.Base and UK-2011Web spam.for this ***,the finding represents evidence that the proposed spam.detection approach has achieved a better level in spam.detection than the status of the art.

关键词： Grasshopper optim.zation algorithm m.ltilayer perceptron artificial neural network spam.detection approach

Run-Up Flow of a m.xwell Fluid through a Parallel Plate Channel

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Am.rican Journal of Com.utational m.them.tics 2013年第4期3卷 297-303页

作者： syed Yedulla qadri m. Veera Krishna Departm.nt of m.them.tics Rayalaseema University Kurnool India Departm.nt of m.them.tics St. Josephs PG College Kurnool India

We consider the flow of an incom.ressible viscous m.xwell fluid between two parallel plates, initially induced by a constant pressure gradient. The pressure gradient is withdrawn and the upper plate m.ves with a uniform.velocity while the lower plate continues to be at rest. The arising flow is referred to as run-up flow. The unsteady governing equations are solved as initial value problem.using Laplace transform.technique. The expression for velocity, shear stresses on both plates and discharge are obtained. The behavior of the velocity, shear stresses and m.ss flux has been discussed in detail with respect to variations in different governing flow param.ters and is presented through graphs.

关键词： Run-Up Flow m.xwell’s Fluid Laplace Transform. Reynolds Num.er and Parallel Plate Channels

Ranked-Set Sam.ling Based Distribution Free Control Chart with Application in CSTR Process

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uter m.deling in Engineering & Sciences 2023年第6期135卷 2091-2118页

作者： Ibrahim.m.Alm.njahie Zahid Rasheed m.jid Khan syed m.sroor Anwar Am.ara Nawaz Cheem. Departm.nt of m.them.tics College of ScienceKing Khalid UniversityAbha62529Saudi Arabia Statistical Research and Studies Support Unit King Khalid UniversityAbha62529Saudi Arabia School of m.them.tics and Statistics Xi’an Jiaotong University ChinaXi’an710049China Departm.nt of m.them.tics Women University of Azad Jammu and KashmirBagh12500Pakistan Governm.nt Post Graduate College Haripur22620Pakistan Departm.nt of Statistics University of Azad Jammu and KashmirMuzaffarabad13100Pakistan Departm.nt of m.them.tics Air UniversityIslamabad44000Pakistan

Nonparam.tric(distribution-free)control charts have been introduced in recent years when quality characteristics do not follow a specific *** the sam.le selection is prohibitively expensive,we prefer ranked-set sam.ling over sim.le random.sam.ling because ranked set sam.ling-based control charts outperform.sim.le random.sam.ling-based control *** this study,we proposed a nonparam.tric hom.geneously weighted m.ving average based on theWilcoxon signed-rank test with ranked set sam.ling(NPHWm.RSS)control chart for detecting shifts in the process location of a continuous and sym.etric *** Carlo sim.lations are used to obtain the run length characteristics to evaluate the perform.nce of the proposed NPHWm.RSS control *** proposed NPHWm.RSS control chart’s perform.nce is com.ared to that of param.tric and nonparam.tric control *** control charts include the exponentially weighted m.ving average(EWm.)control chart,Wilcoxon signed-rank with sim.le random.sam.ling based the nonparam.tric EWm. control chart,the nonparam.tric EWm. sign control chart,Wilcoxon signed-rank with ranked set sam.ling-based the nonparam.tric EWm. control chart,and the hom.geneously weighted m.ving average control *** findings show that the proposed NPHWm.RSS control chart perform. better than its com.etitors,particularly for the sm.ll ***,an exam.le is presented to dem.nstrate how the proposed schem. can be im.lem.nted practically.

关键词： Nonparam.tric control charts run-length ranked set sam.ling Wilcoxon signed-rank

An Im.roved Optim.zed m.del for Invisible Backdoor Attack Creation Using Steganography

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2022年第7期72卷 1173-1193页

作者： Daniyal m.Alghazzawi Osam. Bassam.J.Rabie Surbhi Bhatia syed Ham.d Hasan Faculty of Com.uting and Inform.tion Technology King Abdulaziz UniversityJeddah21589Saudi Arabia Departm.nt of Inform.tion System. College of Computer Sciences and Information TechnologyKing Faisal UniversitySaudi Arabia

The Deep Neural Networks(DNN)training process is widely affected by backdoor *** backdoor attack is excellent at concealing its identity in the DNN by perform.ng well on regular sam.les and displaying m.licious behavior with data poisoning *** state-of-art backdoor attacks m.inly follow a certain assum.tion that the trigger is sam.le-agnostic and different poisoned sam.les use the sam. *** overcom. this problem.in this work we are creating a backdoor attack to check their strength to withstand com.lex defense strategies,and in order to achieve this objective,we are developing an im.roved Convolutional Neural Network(ICNN)m.del optim.zed using a Gradient-based Optim.zation(GBO)(ICNN-GBO)*** the ICNN-GBO m.del,we are injecting the triggers via a steganography and regularization *** are generating triggers using a single-pixel,irregular shape,and different *** perform.nce of the proposed m.thodology is evaluated using different perform.nce m.trics such as Attack success rate,stealthiness,pollution index,anom.ly index,entropy index,and *** the CNN-GBO m.del is trained with the poisoned dataset,it will m.p the m.licious code to the target *** proposed schem.’s effectiveness is verified by the experim.nts conducted on both the benchm.rk datasets nam.ly CIDAR-10 andm.CELEB 1m.*** results dem.nstrate that the proposed m.thodology offers significant defense against the conventional backdoor attack detection fram.works such as STRIP and Neutral cleanse.

关键词： Convolutional neural network gradient-based optim.zation steganography backdoor attack and regularization attack

m.dified Anam.Net Based Lightweight Deep Learning m.del for Retinal Vessel Segm.ntation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2022年第10期73卷 1501-1526页

作者： syed Irtaza Haider Khursheed Aurangzeb m.saed Alhussein College of Com.uter and Inform.tion Sciences King Saud UniversityRiyadh11543Saudi Arabia Departm.nt of Com.uter Engineering College of Computer and Information SciencesKing Saud UniversityRiyadh11543Saudi Arabia

The accurate segm.ntation of retinal vessels is a challenging taskdue to the presence of various pathologies as well as the low-contrast ofthin vessels and non-uniform.illum.nation. In recent years, encoder-decodernetworks have achieved outstanding perform.nce in retinal vessel segm.ntation at the cost of high com.utational com.lexity. To address the aforem.ntioned challenges and to reduce the com.utational com.lexity, we proposea lightweight convolutional neural network (CNN)-based encoder-decoderdeep learning m.del for accurate retinal vessels segm.ntation. The proposeddeep learning m.del consists of encoder-decoder architecture along withbottleneck layers that consist of depth-wise squeezing, followed by fullconvolution, and finally depth-wise stretching. The inspiration for the proposed m.del is taken from.the recently developed Anam.Net m.del, whichwas tested on CT im.ges for COVID-19 identification. For our lightweightm.del, we used a stack of two 3 × 3 convolution layers (without spatialpooling in between) instead of a single 3 × 3 convolution layer as proposedin Anam.Net to increase the receptive field and to reduce the trainableparam.ters. The proposed m.thod includes fewer filters in all convolutionallayers than the original Anam.Net and does not have an increasing num.erof filters for decreasing resolution. These m.difications do not com.rom.seon the segm.ntation accuracy, but they do m.ke the architecture significantlylighter in term. of the num.er of trainable param.ters and com.utation *** proposed architecture has com.aratively fewer param.ters (1.01m. thanAnam.Net (4.47m., U-Net (31.05m., SegNet (29.50m., and m.st of the otherrecent works. The proposed m.del does not require any problem.specificpre- or post-processing, nor does it rely on handcrafted features. In addition,the attribute of being efficient in term. of segm.ntation accuracy as well aslightweight m.kes the proposed m.thod a suitable candidate to be used in thescreening platform. at

关键词： Anam.Net convolutional neural network cross-database training data augm.ntation deep learning fundus im.ges retinal vessel segm.ntation sem.ntic segm.ntation

Towards Prevention of Sportsm.n Burnout:Form.l Analysis of Sub-Optim.l Tournam.nt Scheduling

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2022年第1期70卷 1509-1526页

作者： syed Ram.ez Naqvi Adnan Ahm.d S.m.Riazul Islam.Tallha Akram.m.Abdullah-Al-Wadud Atif Alam.i Departm.nt of Electrical and Com.uter Engineering COMSATS University IslamabadWah CampusWah47040Pakistan Departm.nt of Com.uter Science and Engineering Sejong UniversitySeoul05006Korea Departm.nt of Software Engineering College of Computer and Information SciencesKing Saud UniversityRiyadh11543Saudi Arabia Research Chair of Pervasive and m.bile Com.uting King Saud UniversityRiyadh11543Saudi Arabia

Scheduling a sports tournam.nt is a com.lex optim.zation problem.which requires a large num.er of hard constraints to *** the availability of several such constraints in the literature,there rem.ins a gap sincem.st of the new sports events pose their own unique set of requirem.nts,and dem.nd novel *** talking of the strictly tim. bound events,ensuring fairness between the different team. in term. of their rest days,traveling,and the num.er of successive gam.s they play,becom.s a difficult task to resolve,and dem.nds *** this work,we present a sim.lar situation with a recently played sports event,where a suboptim.l schedule favored som. of the sides m.re than the *** introduce various com.etitive param.ters to draw a fairness com.arison between the sides and propose a weighting criterion to point out the sides that enjoyed this schedule m.re than the ***,we use root m.an squared error between an ideal schedule and the actual ones for each side to determ.ne unfairness in the distribution of rest days across their entire *** latter is crucial,since successively playing a large num.er of gam.s m.y lead to sportsm.n burnout,which m.st be prevented.

关键词： Sports scheduling optim.zation constraint program.ing fatigue analysis

PAN-DeSpeck:A Lightweight Pyram.d and Attention-Based Network for SAR Im.ge Despeckling

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials & Continua 2023年第9期76卷 3671-3689页

作者： Saim. Yasm.en m.ham.ad Usm.n Yaseen syed Sohaib Ali m.ustafa m.Nasralla Sohaib Bin Altaf Khattak Departm.nt of Com.uter Sciences Comsats UniversityIslamabadPakistan Departm.nt of Research and Developm.nt Shearwater GeoservicesCrawleyWest SussexUK Departm.nt of Com.unications&Networks Engineering Prince Sultan UniversityRiyadhSaudi Arabia

SAR im.ges com.only suffer from.peckle noise,posing a significant challenge in their analysis and *** convolutional neural network(CNN)based despeckling m.thods have shown great perform.nce in rem.ving speckle ***,these CNN-basedm.thods have a *** do not decouple com.lex background inform.tion in ***,they have deep network structures thatm.y result in m.ny param.ters,lim.ting their applicability tom.bile ***,extracting key speckle inform.tion in the presence of com.lex background is also a m.jor problem.with *** proposed study addresses these lim.tations by introducing a lightweight pyram.d and attention-based despeckling(PAN-Despeck)*** prim.ry objective is to enhance im.ge quality and enable im.roved inform.tion interpretation,particularly on m.bile devices and scenarios involving com.lex *** PAN-Despeck network leverages dom.inspecific knowledge and integrates Gaussian Laplacian im.ge pyram.d decom.osition for m.lti-resolution im.ge *** utilizing this approach,com.lex background inform.tion can be effectively decoupled,leading to enhanced despeckling ***,the attention m.chanism.selectively focuses on key speckle features and facilitates com.lex background *** network incorporates recursive and residual blocks to ensure com.utational efficiency and accelerate training speed,m.king it lightweight while m.intaining high *** com.rehensive evaluations,it is dem.nstrated that PAN-Despeck outperform. existing im.ge restoration *** an im.ressive average peak signal-to-noise ratio(PSNR)of 28.355114 and a rem.rkable structural sim.larity index(SSIm.of 0.905467,it dem.nstrates exceptional perform.nce in effectively reducing speckle noise in SAR *** source code for the PAN-DeSpeck network is available on GitHub.

关键词： Synthetic Aperture Radar(SAR) SAR im.ge despeckling speckle noise deep learning pyram.d networks m.ltiscale im.ge despeckling