Gear Fault Diagnosis Based on Rough Set and Support Vector Machine

By introducing Rough Set Theory and the principle of Support vector machine,a gear fault diagnosis method based on them is proposed.Firstly,diagnostic decision-making is reduced based on rough set theory,and the noise and redundancy in the sample are removed,then,according to the chosen reduction,a support vector machine multi-classifier is designed for gear fault diagnosis.Therefore,SVM’training data can be reduced and running speed can quicken.Test shows its accuracy and effi- ciency of gear fault diagnosis.

On the Differentiability of Vector Valued Additive Set Functions

The Lebesgue-Nikodym Theorem states that for a Lebesgue measure an additive set function ?which is -absolutely continuous is the integral of a Lebegsue integrable a measurable function;that is, for all measurable sets.?Such a property is not shared by vector valued set functions. We introduce a suitable definition of the integral that will extend the above property to the vector valued case in its full generality. We also discuss a further extension of the Fundamental Theorem of Calculus for additive set functions with values in an infinite dimensional normed space.

EKELAND'S PRINCIPLE FOR SET-VALUED VECTOR EQUILIBRIUM PROBLEMS

In this paper, we introduce a concept of quasi C-lower semicontinuity for setvalued mapping and provide a vector version of Ekeland＇s theorem related to set-valued vector equilibrium problems. As applications, we derive an existence theorem of weakly efficient solution for set-valued vector equilibrium problems without the assumption of convexity of the constraint set and the assumptions of convexity and monotonicity of the set-valued mapping. We also obtain an existence theorem of ε-approximate solution for set-valued vector equilibrium problems without the assumptions of compactness and convexity of the constraint set.

Benson proper efficiency for vector optimization of generalized subconvexlike set-valued maps in ordered linear spaces

Several equivalent statements of generalized subconvexlike set-valued map are established in ordered linear spaces. Using vector closure, we introduce Benson proper efficient solution of vector optimization problem. Under the assumption of generalized subconvexlikeness, scalarization, multiplier and saddle point theorems are obtained in the sense of Benson proper efficiency.

Construction and expression of SET gene and siRNA recombinant adenovirus vectors

Objective:To construct SET gene recombinant adenovirus vector and SET gene small interfering RNA(SiRNA) recombinant adenovirus vector for over-expression or knock-down of SET levels. Methods:The cDNA sequence of SET was cloned by reverse transcriptive polymerase chain reaction(RT-PCR) and the SET gene fragment was subcloned into adenovirus shuttle plasmid pAdTrack-CMV to construct the shuttle plasmid pAdTrack-SET.The shuttle plasmid pAdtrack-SET was transformed into BJ5183 cells with the adenoviral backbone pAdEasy-1 to obtain the homologous recombinant Ad-CMV-SET and the recombinant Ad-CMV-SET was packaged and amplified in the AD293 cells.The expression of SET in AD293 cells was detected by Western blot.In addition,we constructed SET gene SiRNA recombinant adenovirus vector(Ad-H1-SiRNA/SET) and its efficacy of knockdown of SET protein was detected in infected GC-2spd(ts) cells by Western blot. Results:The recombinant adenovirus vectors,both SET gene recombinant adenovirus vector Ad-CMV-SET and SET gene SiRNA recombinant adenovirus vector Ad-H1-SiRNA/SET,were proven to be constructed successfully by the evidence of endonulease digestion and sequencing.AD293 cells infected with either recombinant adenovirus vector of Ad-CMV-SET or Ad-H1-SiRNA/SET were observed to express GFP.The expression of SET protein was up-regulated significantly in AD293 cells infected with SET gene recombinant adenovirus vector.On the contrast, SET protein was significantly down-regulated in the GC-2spd(ts) cells infected with Ad-H1-SiRNA/SET (P<0.05) and the knockdown efficiency was approximately 50%-70%. Conclusion:The recombinant adenovirus vector Ad-CMV-SET and Ad-H1-SiRNA/SET were successfully constructed and effectively expressed in germ cells and somatic cells.It provides an experimental tool for further study of SET gene in the physiological and pathophysiological mechanism of reproduction-related diseases.

Optimality for Henig Proper Efficiency in Vector Optimization Involving Dini Set-Valued Directional Derivatives

This note studies the optimality conditions of vector optimization problems involving generalized convexity in locally convex spaces. Based upon the concept of Dini set-valued directional derivatives, the necessary and sufficient optimality conditions are established for Henig proper and strong minimal solutions respectively in generalized preinvex vector optimization problems.

Upper and Lower Semicontinuity of Solution Sets for Parametric Generalized Vector Quasi-equilibrium Problems

In this paper, two kinds of parametric generalized vector quasi-equilibrium problems are introduced and the relations between them are studied. The upper and lower semicontinuity of their solution sets to parameters are investigated.

Gap Functions and Error Bounds for Set-Valued Vector Quasi Variational Inequality Problems

One of the classical approaches in the analysis of a variational inequality problem is to transform it into an equivalent optimization problem via the notion of gap function. The gap functions are useful tools in deriving the error bounds which provide an estimated distance between a specific point and the exact solution of variational inequality problem. In this paper, we follow a similar approach for set-valued vector quasi variational inequality problems and define the gap functions based on scalarization scheme as well as the one with no scalar parameter. The error bounds results are obtained under fixed point symmetric and locally α-Holder assumptions on the set-valued map describing the domain of solution space of a set-valued vector quasi variational inequality problem.

Why Are There as Many Elements in the Cantor Set as There Are Real Numbers?

There are many important concepts in linear algebra, such as linear correlation and linear independence, eigenvalues and eigenvectors, and so on. The article provides a graphical explanation of how to distinguish between the concepts of linear correlation and linear independence. The conclusion points out that linear independence means that there are no two (base) vectors with the same direction in a vector graph;otherwise, it is a linear correlation.

基于决策树的社交网络隐式用户行为数据挖掘方法

为了解决社交网络隐式用户行为数据挖掘过程中关联相似性计算较为困难的问题,提出了基于决策树的社交网络隐式用户行为数据挖掘方法。将社交网络视为包含不同维度的向量空间,计算特定维度上用户的兴趣空间和兴趣点。确定样本属性集后,根据已知行为数据建立测试分支,计算该分支下子集的属性权重,不断迭代直至挖掘到同等属性的数据点为止。测试结果表明:该方法可对不同种类隐式用户行为精准挖掘,目标行为数据查找效果较好,实用性较强。

基于BiLSTM的NL2SQL模型

随着互联网技术的发展,很多应用为大众提供金融量化服务,而大部分用户不具备金融或计算机专业知识,他们期望使用自然语言查询数据,因此自然语言转SQL(NL2SQL)被迫切需要。针对此问题,提出一种基于双向长短期记忆模型(BiLSTM)的中文金融NL2SQL算法,分为编码和解码阶段。在编码阶段,利用BiLSTM和注意力机制生成特征向量。在解码阶段,根据SQL的语法规则,将SQL生成解耦为九个分类任务,各个任务间相互依赖联合学习,之后生成复杂的SQL语句。除模型外,还训练出包含金融词汇的向量库,构建金融领域的数据集。通过在此数据集上实验验证,结果表明,该方法准确率更高,能有效解决金融领域SQL生成问题,并在某金融量化分析系统中实现。

SVM样本约简算法研究综述

支持向量机(Support Vector Machine, SVM)是基于统计学习理论和结构风险最小化原则发展起来的一种有监督的机器学习算法,它有效克服了局部最小和维数灾难等问题,具有良好的泛化性能,并被广泛应用于模式识别和人工智能领域。但SVM的学习效率随着训练样本数量的增加而显著降低,对于大规模训练集,采用标准优化方法的传统SVM面临着内存需求过大、执行速度慢,有时甚至无法执行的问题。为了缓解SVM在大规模训练集上存储需求高、训练时间长等问题,学者们提出了SVM样本约简算法。文中首先介绍了SVM理论基础,然后从基于聚类、几何分析、主动学习、增量学习和随机抽样5个方面系统综述了SVM样本约简算法的研究现状,讨论了各种SVM样本约简算法的优缺点,最后总结全文并展望未来。

人SET基因小分子干扰RNA重组腺病毒载体的构建与鉴定

目的SET基因表达产物是SET／TAF-1β蛋白，功能涉及基因表达调控、染色质修饰、翻译后修饰等生物过程，多水平、多通路地调节靶因子，其表达异常与多种疾病相关。为研究SET基因与临床多种疾病发生发展的关系，文中构建表达人SET基因的小分子干扰RNA（siRNA）重组腺病毒载体。方法根据siRNA靶序列设计并合成2条互补的64 bp寡核苷酸。pShuttle-H1穿梭质粒经BglⅡ、HindⅢ双酶切后，与退火的寡核苷酸连接，构建穿梭质粒pShuttle-H1-siRNA。该穿梭质粒经Not Ⅰ和Hind Ⅲ双酶切后与pAdTrack-CMV连接，构建含有绿色荧光蛋白（green fluorescense protein，GFP）报告基因的穿梭质粒PATC—H1-siRNA。分别将腺病毒骨架质粒pAdEasy-1和穿梭质粒PATC—H1-siRNA转化至BJ-5183感受态细菌中进行同源重组。Pac Ⅰ酶切线性化重组质粒pAd—H1-siRNA后，转染AD293细胞进行病毒包装和扩增。根据GFP报告基因的表达观察重组病毒的产生和感染效率，用Western blot检测人SET蛋白的表达水平。结果穿梭质粒pShuttle-H1-siRNA经双酶切和测序证实构建成功；重组腺病毒载体经AD293细胞包装后可观察到GFP表达；获得的重组腺病毒载体感染AD293细胞。经Western blot检测，SET基因腺病毒载体（AdH1-SiRNA／SET）感染组与未感染腺病毒载体组（空白对照）和感染对照腺病毒载体组（AdH1-SiRNA／NS）相比，SET蛋白表达水平显著降低（P〈0．05），抑制效率为50％～70％。结论成功构建了表达人SET基因的siRNA重组腺病毒载体AdH1-siRNA／SET，并在AD293细胞中验证其抑制SET蛋白表达的作用，为进一步研究SET基因参与多种疾病的发生发展提供了材料。

基于Level Set方法的Visible Human Being虚拟人图像处理

提出了将基于区域的VectorConfidenceConnected的低级分割方法和基于边界的LevelSet高级分割方法相组合的分割方法,使得边界的平滑性得到了保持;既有效地结合了医学专家的医学背景知识,又提高了分割处理的速度,很好的处理了医学图像中常见的拓扑结构,实验结果表明该方法切实可行。

基于特征约简与改进支持向量机的动态过程质量异常识别方法

为了有效降低特征维数并提高动态过程异常模式的识别精度,提出基于特征约简与改进支持向量机的动态过程质量异常识别方法。本文首先提取能反映质量异常模式的16个统计特征与7个形状特征,再使用粗糙集(RS)约简特征集合以剔除冗余特征与干扰特征。同时,使用遗传算法(GA)寻找支持向量机(SVM)的最优参数,并采用GA-SVM模型识别质量异常模式。仿真实验表明:粗糙集筛选后得到的12个特征具有较强区分动态过程是否出现异常状况的能力,遗传算法参数寻优后的支持向量机识别质量异常模式的精度明显高于其他同类型的模型,因此,本文提出的RS-GA-SVM模型具有良好的识别精度与稳健性,能够对动态过程进行有效监控。

SET基因shRNA慢病毒载体的构建及其在肝细胞中的表达鉴定

目的:构建SET基因shRNA慢病毒表达载体,为研究SET在三氯乙烯毒性机制中的作用提供技术支持。方法:通过NCBI检索SET序列,设计合成5对shRNA片段,退火后连接到慢病毒载体pLVX-shRNA1 vector。挑取筛选的单菌落,经PCR和测序鉴定后提取质粒。将质粒共转染到293T细胞,收集病毒上清,转导入L-02肝细胞中,利用嘌呤霉素筛选得到SET缺陷型细胞,最后利用荧光定量PCR和Western blot鉴定干扰效果。结果:PCR和测序结果证明双链shRNA正确插入慢病毒载体pLVX-shRNA1 vector,共转染293T细胞得到高滴度病毒,转导L-02细胞筛选出SET基因缺陷型细胞。结论:利用慢病毒介导的RNAi技术成功构建了SET基因缺陷型细胞。

基于改进蜻蜓算法的磁齿轮复合电机控制研究

针对磁齿轮复合电机双转子结构在无接触传动时出现振荡和超调的问题,提出了一种基于改进蜻蜓算法(improved dragonfly algorithm,IDA)的双闭环PI(proportional integral,比例积分)参数自整定矢量控制方法。针对DA在收敛速度和收敛精度等方面存在的不足,分别在算法寻优的前期、中期和后期引入Tent映射、改进权重系数和差分优化算法,并在其适应度函数上增加可抑制振荡和超调的惩罚项,使算法的收敛速度和收敛精度得到明显提高。采用PI、DA-PI和IDA-PI三种控制方法对磁齿轮复合电机进行控制仿真和实验,结果表明,在IDA-PI控制下电机转速的超调量和稳态误差最小,动态响应速度最快,证明了所提策略的有效性。研究结果为不同拓扑结构磁齿轮复合电机的控制提供了参考。

开集环境中基于增量学习的网络流量分类研究

面对网络流量新类别不断涌现的挑战,以及随之而来的开集识别和模型更新需求,文章提出了一种基于增量学习的开集网络流量分类方法。对于开集识别,支持向量机和K均值聚类算法的级联结构可以持续识别新类和已知类;对于模型更新,基于候选支持向量筛选的“样本回放”和新旧模型加权融合的“参数回放”方法,能有效解决“有类增量的灾难性遗忘”问题。与ISK和DACS方法相比,该方法应用在开集流量识别和分类任务中表现出显著优势,F1分数能提高1百分点至8百分点,分类速度也优于现有方法。

基于正域向量的决策粗糙集属性约简

在决策粗糙集模型中,现有划分层的正域都是通过集合求并所得,但基于该方法的保正域不变的属性约简与基于差别矩阵方法所得约简结果并不一致。提出了一种基于正域向量的决策粗糙集属性约简方法,该方法与基于差别矩阵的约简方法所得结果是一致的。最后给出一个算例说明其一致性。

基于AVX512的格密码高速并行实现

量子计算的迅速发展可能对当前广泛使用的公钥密码算法造成严重威胁。格密码因优秀的抗量子安全性和高效的计算效率在后量子密码中占据重要地位。美国国家标准技术研究院于2022年5月公布4个后量子密码标准,其中3个是格密码算法,Kyber算法便是其中之一。随着后量子密码标准的确定,Kyber算法高效实现的需求日益增加。基于512位高级向量扩展(AVX512),对Kyber算法进行优化与高速并行实现。使用惰性模约减、优化的蒙哥马利模约减及优化的快速数论变化等技术,充分利用计算机的存储空间,减少大量不必要的模约减操作,提高多项式计算的效率与并行性。采用冗余比特技术,增强多项式抽样过程中比特的并行处理能力。通过AVX512的512 bit位宽和8路并行实现哈希运算,并对其产生的伪随机比特串进行合理调度,充分发挥并行性能。基于AVX512指令集高速并行实现Kyber上的多项式计算和抽样,并进一步实现整个Kyber公钥加密方案。性能测试结果表明,与C语言实现相比,基于AVX512实现的密钥生成和加密算法获得了10~16倍的加速,解密算法获得了约56倍的加速。