癌症基因表达数据的集成分类器设计与分析

基因表达水平对癌症诊断起到重要的作用。文章提出了一种基于SVM(Support Vector Machine)的集成分类算法,从基因表达水平的微阵列数据中对癌症和正常群体进行分类。文章提出了一种结合Adaboost算法和递归特征消除(Recursive Feature Elimination,RFE)算法,选取最显著的特征并设计与之适合的分类器。该方法已应用于癌症病人的基因表达微阵列数据的分类中,其分类结果在成功率方面有极大的提升。     

一种改进的最小二乘孪生支持向量机分类算法

提出了一种新的模式分类器,即广泛权重的最小二乘孪生支持向量机.该支持向量机在正、负两类样本上广泛地增加权重,很好地抑制了交叉噪声样本对数据分类的影响.其次,根据间隔最大化原理,该支持向量机在目标函数上增加了一个正规化项,实现结构风险最小化和避免在求解该目标函数时可能对病态矩阵求逆的处理.同时,提出了利用一种指数函数计算训练样本的密度来获得样本权重值的算法.该算法能够有效缩减计算权重的时间,且具有较强的鲁棒性.实验证明本文提出的广泛权重的最小二乘孪生支持向量机能够实现高精度和高效率的分类效果,而且特别适合于含有交叉噪声样本的数据集分类.     

大数据环境下并行数据流预测及优化方法

以模糊集成联机分析处理(On-Line Analytical Processing,OLAP)为基础对关联规则挖掘算法进行改进,该算法为多维联机研究提供一种关联规则计算架构。基于模糊数据立方体的知识发现,为用户提供灵活的多维数据层次抽象模式。在多维数据集的多维属性处理中,引入模糊数据立方体作为问题措施补充,并利用不同层次的模糊关联规则构造模糊数据立方体,然后利用权重和多层次的概念构建模糊加权多层次关联规则。最后,通过对所提算法在合成数据集以及2000年中国人口普查的数据仿真测试,验证了基于OLAP的挖掘方法要比离散关联规则挖掘方法、单独支持阈值关联规则及最小挖掘项集关联规则三种对比算法,在最小支持度、置信度、权重均值等指标上,性能更加优异。     

基于传播算子的声学矢量传感器阵列 扩展孔径二维DOA估计算法

 本文提出一种基于传播算子的声学矢量传感器阵列扩展孔径二维DOA估计算法.首先,利用传播算子方法得到一组高精度模糊的DOA估计值;然后,利用声学矢量传感器的特点得到另一组低精度无模糊的DOA估计值;最后,利用无模糊估计值对模糊估计值进行解模糊处理,得到高精度无模糊的DOA估计值.提出的算法无需进行特征值分解或奇异值分解进行信号子空间/噪声子空间的估计.与基于ESPRIT的算法相比,提出的算法的计算量约为信号个数与声学矢量传感器个数的四倍之比.计算机仿真结果表明在信噪比不是很低时,提出的算法与基于ESPRIT的算法具有相当的估计性能.     

基于自适应权重马尔科夫场的彩色纹理分割

提出基于自适应权重马尔科夫场的无监督分割模型.利用HSV空间上的彩色信息和Gabor纹理特征组成MRF数据特征分量,通过马尔科夫最大后验概率框架和能量函数最小化算法将像素分类获得分割结果.在自然纹理与合成纹理上的实验证明了此算法的有效性并探讨了4种能量最小化方法的质量.     

基于稀疏表示与特征融合的人脸识别方法

针对人脸识别在有遮挡、表情、光照的变化或受到噪声污染时鲁棒性变差问题,提出一种基于稀疏表示与特征融合的人脸识别算法。首先采用低秩恢复算法得到训练样本和测试样本的干净人脸图像,提取干净人脸图像的LBP,HOG,Gabor三种特征向量;然后对部分训练样本进行SRC分类测试,根据SRC的识别结果与分类残差定义一个损失函数,再利用正则化最小二乘法计算出使损失函数最小的权重向量;最后根据该权重向量重构规则化残差进行分类。在ORL,Extended Yale B和AR数据库上进行实验,结果表明,该算法优于利用单一特征识别的方法,并且对光照、噪声、遮挡等因素产生的影响有较好的泛化性能。     

基于非线性核空间映射与人工免疫网络的高光谱遥感图像分类

提出了一种基于非线性核空间映射人工免疫网络的高光谱遥感图像分类算法.根据生物免疫网络基本原理构建了人工免疫网络模型,利用非线性核函数将高光谱训练样本映射到高维空间,完善了人工免疫网络中目标样本核空间相似性分选方法,降低了人工免疫网络识别样本所需的抗体数量,提升了算法的分类精度和运算效率.为了验证算法的有效性,利用两组高光谱遥感数据将多种高光谱分类方法进行了对比实验.实验表明该算法分类精度和算法运算时间上都有较大改善,是一种分类精度更高、运算速度更快的改进型基于人工免疫网络的高光谱遥感图像分类新方法.     

基于最小二乘隐空间支持向量机的IDS检测算法的设计

在基于支持向量机的基础上,提出一种新的利用最小二乘隐空间支持向量机设计IDS的检测算法,解决了网络入侵检测系统中检测算法的分类精度不高、训练样本数需要较多,及训练学习时间较长等问题.仿真实验结果表明,本算法较基于支持向量机的检测算法具有更快的收敛性、更快的迭代速度、更高的检测精度和更低的误报率.     

基于EP的提升分类算法

显露模式(EP)是支持度从一个数据集到另一个数据集发生显著变化的项集.EP具有很强的区分能力,可以建立很好的分类器.文中采用基于EP的分类算法CEEP建立基分类器,结合组合学习分类方法AdaBoost算法的思想,提出了一种新的分类算法A-E算法.算法使用加权样本建立基分类器,并根据分类结果改变样本权值,同时应用分类误差计算基分类器权重.最终,算法按权重组合每个分类器的分类结果.在UCI机器学习数据库的9个基准数据集上的实验表明,A-E算法都能有效地减低泛化误差,并具有较高的分类准确率.     

基于PLS的加权朴素贝叶斯分类测试算法

朴素贝叶斯算法是一种简单而高效的分类算法,但是它的条件独立性假设影响了其分类性能。通过放松朴素贝叶斯假设,可以增强其分类效果,但通常会导致计算代价大幅提高。文章提出了一种基于偏最小二乘的加权朴素贝叶斯分类算法,通过建立条件属性和决策属性之间偏最小二乘回归方程,把回归系数赋给对应的条件属性,作为相应的权重,从而在保持简单性的基础上有效地提高了朴素贝叶斯算法的分类性能。最后,通过在UCI数据集上的仿真实验,验证了该算法的有效性。     

Adaptive Multi-phenotype Based Gene Expression Programming Algorithm

Expression theory is the mathematical foundation of evolutionary computation.In order to investigate the problems in Gene expression programming (GEP) expression theory,we clarified the difference between genotypic expression space and phenotypic expression space.We also presented phenotypic expression space definition and theory.Then we analyzed the reason of good and bad performance of different GEP algorithms based on expression space theory.We also proposed a new Adaptive multi-phenotype gene expression programming (AMGEP),in which the potential of genes is fully activated with gene combination.Experiments on benchmark problems showed that genotypic expression space and phenotypic expression space theory can explain the different performance of different algorithms and showed that AMGEP outperform other GEP algorithms in terms of search ability.     

复合离散混沌动力系统与序列密码体系

本文构造了复合离散混沌动力系统,并研究了它不变分布和迭代轨迹的若干性质.利用两个特殊的离散混沌系统,提出了基于复合离散混沌动力系统的序列密码体系.由于复合离散混沌系统对初始条件的敏感性和迭代过程的伪随机性,本文提出的序列密码算法将明文和密钥序列完全融合在密文序列中,使它们与密文之间形成了复杂而敏感的非线性关系,这就防止了密文对有关信息的泄露,从而使系统具有很高的安全性.该算法还拥有很大的密钥空间和均匀分布的密文.     

基于谱回归和核空间最近邻的基因表达数据分类

肿瘤基因表达数据是典型的高维小样本数据,直接对其进行识别存在维数灾难,需要对数据进行维数约简.提出了一种基于谱回归分析和核空间最近邻分类器的基因表达数据分类方法,采用谱回归分析得到可有效提取低维鉴别特征的投影矩阵,然后通过投影矩阵对基因表达数据进行维数约简,得到的低维数据用核空间最近邻分类器进行识别.通过在Prosta...     

基于跨视角相似度顺序保持的基因特征提取方法

基因表达数据通常具有维数高、样本少、类别分布不均等特点,如何提取基因表达数据的有效特征是基因分类研究的关键问题。该文借助相关分析理论,构建鉴别敏感的视角内相似度顺序保持散布并且约束鉴别敏感的视角间相似度相关,从而形成了一种新的基因特征提取方法,即相似度顺序保持跨视角相关分析(SOPACA)。该文方法在保持不同视角间特征类内聚集性和相似度顺序的同时具有较大的类间离散性。在癌症基因表达数据集上的良好实验结果显示了该文方法的有效性。     

Improving reliability of gene selection from microarray functional genomics data

Constructing a classifier based on microarray gene expression data has recently emerged as an important problem for cancer classification. Recent results have suggested the feasibility of constructing such a classifier with reasonable predictive accuracy under the circumstance where only a small number of cancer tissue samples of known type are available. Difficulty arises from the fact that each sample contains the expression data of a vast number of genes and these genes may interact with one another. Selection of a small number of critical genes is fundamental to correctly analyze the otherwise overwhelming data. It is essential to use a multivariate approach for capturing the correlated structure in the data. However, the curse of dimensionality leads to the concern about the reliability of selected genes. Here, we present a new gene selection method in which error and repeatability of selected genes are assessed within the context of M-fold cross-validation. In particular, we show that the method is able to identify source variables underlying data generation.     

基于小波包-FastICA在阿尔茨海默症中的应用及其生物学分析

阿尔茨海默症(Alzheimer’s disease,AD)基因表达谱数据具有高维性、高噪声、高冗余性等特点,使得AD特异性基因的搜索空间巨大,搜索算法时间长,降低了算法的挖掘性能及其生物学分析。因此对其基因表达谱数据进行去噪和降维预处理是十分必要的。文中首先利用小波包变换-SAM方法对数据进行降维去噪,实验结果证明了小波包方法能较好地提取基因表达谱有用信息;然后应用快速独立成分分析(FastICA)算法对预处理后的数据进行矩阵分解分析,并根据独立分量选取特异性基因。在此基础上的样本分类实验表明,FastICA提取的特异性基因具有较高的显著性,能够提高样本的分类结果。同时,通过所提取特异性基因的富集性分析,文中给出了这些基因在阿尔茨海默症数据集中聚类情况及其基因表达情况,为AD的生物学及医学病理分析提供有利的依据。     

A modified binary particle swarm optimization for selecting the small subset of informative genes from gene expression data

Gene expression data are expected to be of significant help in the development of efficient cancer diagnoses and classification platforms. In order to select a small subset of informative genes from the data for cancer classification, recently, many researchers are analyzing gene expression data using various computational intelligence methods. However, due to the small number of samples compared to the huge number of genes (high dimension), irrelevant genes, and noisy genes, many of the computational methods face difficulties to select the small subset. Thus, we propose an improved (modified) binary particle swarm optimization to select the small subset of informative genes that is relevant for the cancer classification. In this proposed method, we introduce particles' speed for giving the rate at which a particle changes its position, and we propose a rule for updating particle's positions. By performing experiments on ten different gene expression datasets, we have found that the performance of the proposed method is superior to other previous related works, including the conventional version of binary particle swarm optimization (BPSO) in terms of classification accuracy and the number of selected genes. The proposed method also produces lower running times compared to BPSO.     

Tumor Intracellular‐Environment Responsive Materials Shielded Nano‐Complexes for Highly Efficient Light‐Triggered Gene Delivery without Cargo Gene Damage

Gene therapy has great potential to bring tremendous improvement to cancer therapy. Recently, photochemical internalization (PCI) has provided the opportunity to overcome endo‐lysosomal sequestration, which is one of the main bottlenecks in both gene and chemotherapeutic delivery. Despite PCI having shown great potential in gene delivery systems, it still remains difficult to perform due to the photo‐oxidation of exogenous cargo genes by reactive oxygen species (ROS) generated from activated photosensitizers (PSs). In this paper, a new type of a stable light‐triggered gene delivery system is demonstrated based on endo‐lysosomal pH‐responsive polymeric PSs, which serve as shielding material for the polymer/gene complex. By taking advantage of the endo‐lysosomal pH‐sensitive de‐shielding ability of the pH‐responsive shielding material incorporated in the ternary gene complexes (pH‐TCs), a more significant photo‐triggered gene expression effect is achieved without damage to the gene from ROS. In contrast, pH‐insensitive material‐shielded nanocarriers cause photo‐oxidation of the payload and do not generate a notable transfection efficacy. Importantly, with the benefit of our newly developed gene delivery system, the deep penetration issue can be resolved. Finally, the light‐triggered gene delivery system using pH‐TCs is applied to deliver the therapeutic p53 gene in melanoma K‐1735 bearing mice, showing excellent therapeutic potential for cancer.     

Highly Efficient Multifunctional Supramolecular Gene Carrier System Self‐Assembled from Redox‐Sensitive and Zwitterionic Polymer Blocks

It has been a challenge to incorporate multiple features into a single gene carrier system to overcome numerous hurdles during the gene delivery. Herein, a supramolecular approach for building a multifunctional gene carrier system is demonstrated with the functions of disulfide bond based reduction‐responsive degradation and zwitterionic phosphorylcholine based extracellular stabilization and favorable cellular uptake. The gene carrier system is self‐assembled from two molecular building blocks: one host polymer, which is a redox‐sensitive β‐cyclodextrin based cationic star polymer, and one guest polymer, which is adamantyl end capped zwitterionic phosphorylcholine based polymer. The host and guest polymers self‐assemble to integrate multiple functions into one system, based on the host‐guest interaction between β‐cyclodextrin and adamantyl moieties. With the rational designs of both building blocks, the supramolecular gene carrier system possesses excellent protein stability, serum tolerance, cellular uptake and intracellular DNA release properties, and also low cytotoxicity. These features work simultaneously to achieve exceptionally high gene transfection efficiency, which is proven in MCF‐7 cell cultures using luciferase and green fluorescence protein reporter genes. Finally, the supramolecular gene carrier is applied to deliver the therapeutic p53 anti‐cancer gene in MCF‐7 cells, showing great potential for cancer gene therapy application.     

粗糙集近似集不确定性研究

粗糙集用上、下近似集刻画不确定目标集合,而粗糙集的近似集用0.5-近似集作为不确定目标集合的近似集.本文首先分析了基于粗糙集的0.5-近似集相似度的属性约简算法存在理论不完备的不足,指出这种相似度具有随知识粒度变化不敏感的缺陷.然后进一步给出了多粒度知识空间下相似度的变化规律,提出了粗糙集近似集的模糊度概念,分析了粗糙集近似集的模糊度在多粒度知识空间下的变化规律,进而提出了相应的属性约简算法.从新的视角构建了目标概念与其近似集的差异性度量方法.