基于最大均值差异的多模态过程过渡模态识别方法

为了更好地揭示多模态过程的运行状态和数据分布变化规律,提高后续建模精度,提出基于最大均值差异（MMD）的多模态过程的过渡模态识别方法. 引入滑动窗口对数据进行切割,使用最大均值差异对多模态过程数据的分布差异进行度量,通过与稳定模态阈值α比较区分过程数据的稳定模态和过渡模态. 在过渡模态段内减小滑动窗口窗宽,利用过渡模态阈值β识别出过渡子模态. 数值仿真实验的模态识别结果表明,所提方法可以准确检测出输入变量期望值的阶跃变化,实现对模态的准确识别. 田纳西伊斯曼（TE）过程仿真数据实验表明,所提方法可以有效地划分出合理的模态,进而选择出分布最相近的历史模态建模,提高多模态过程的软测量建模精度.     

基于小波分解和最大互信息的图像配准方法研究

研究了基于最大互信息图像配准的理论和实现算法,提出了将小波分解与之相结合的算法,证明了该算法优于基于最大互信息的图像配准。通过实验数据比较可以得出,提出的基于小波分解和最大互信息的图像配准的算法可以有效地提高配准的速度和具有较好的配准效果。     

住宅建筑能耗基准确定及用能评价新方法

针对当前建筑能耗基准确定及用能评价方法仍不完善的现状,提出了一种适用于住宅建筑的新方法。通过灰色关联分析确定了不同影响因素(即特征参数)与建筑能耗的关联度,将关联度作为该因素的权值并结合聚类分析对建筑进行合理分类。在此基础上采用累积频率分布法确定了每类建筑的能耗基准值,并对建筑住户进行用能评价。为验证该方法的可行性,将其应用于日本建筑学会所建立的住宅建筑能耗数据库,为聚类后的每类建筑确定了相应的能耗基准并对住户进行了用能评价。结果表明该方法能够综合考虑不同特征参数的影响大小,科学合理地对建筑细化分类,根据分类后的基准评价能够更好地评估住户的节能潜力并提供可行的节能建议。     

敏捷化AIOps框架及运维数据质量评估方法

敏捷化智能运维(AIOps)框架将模型构建提前至测试阶段,利用该阶段产生的监控数据代替线上采集的数据以训练AIOps模型,进而实现智能运维的早开发与早使用. 运维数据质量评估方法通过最大均值差异度量方式,分别在健康评估与故障诊断运维场景下对训练数据分别进行趋势性、阶段性、可检测性及可诊断性评估,以预估数据对模型的适用性. 基于华为技术有限公司提供的测试环境设置测试用例并构建实验数据集,在该数据集上的实验结果验证了敏捷化AIOps框架的可行性及数据质量评估方法的有效性.     

Single channel blind source separation based on ICA feature extraction

A new technique is proposed to solve the blind source separation (BSS) given only a single channel observation. The basis functions and the density of the coefficients of source signals learned by ICA are used as the prior knowledge. Based on the learned prior information the learning rules of single channel BSS are presented by maximizing the joint log likelihood of the mixed sources to obtain source signals from single observation, in which the posterior density of the given measurements is maximized. The experimental results exhibit a successful separation performance for mixtures of speech and music signals.     

基于算术调和均值距离测度的弹性图像配准

针对图像配准中的特征提取问题,提出了一种基于自动选取标记点的弹性图像配准方法。首先,提出了新的相似性测度,算术-调和均值距离测度AHM,图像通过最大化AHM全局配准,然后将图像划分成均匀子块,采用AHM测度刚性配准各对应子块,选取对应子块的中心作为标记点对,最后,用三次均匀B样条实现图像的弹性配准。实验结果表明,AHM是有效的,其配准速度是互信息的2倍。     

基于边缘检测和最大互信息的医学图像配准算法

在临床上,使用多种成像技术并适当地将其加以融合,可为临床诊断和手术治疗提供更加全面准确的信息。而图像配准是融合技术中需要先期解决的问题,也是融合的关键部分。本文提出了一种基于DSP的医学图像配准问题的研究与实现的方案。该方案首先采用TI公司的高性能多媒体数字信号处理器DM642获取实时医学图像,再用配准算法验证其获取的实时医学图像配准的可行性。并采取了基于轮廓提取和最大互信息的方法进行图像配准,克服了互信息单一的利用图像灰度信息的局限性,将边缘检测与互信息相结合,较准确地完成图像配准任务。     

一种基于互信息的波形优化设计方法

将注水法改进成一种适用于杂波背景下的通用注水法,通过最大化目标回波与扩展目标特性之间的互信息降低目标特性的不确定性．然后把这种思想应用于目标识别波形优化问题,从而增加各类目标的可分性．仿真结果证明,相对于线性调频信号和注水信号,该方法优化波形可以得到更大的互信息,并能将目标识别率提高3％左右,在信杂噪比较低的时候提高更加明显．     

Feature selection algorithm for text classification based on improved mutual information

In order to solve the poor performance in text classification when using traditional formula of mutual information (MI),a feature selection algorithm were proposed based on improved mutual information.The improved mutual information algorithm,which is on the basis of traditional improved mutual information methods that enhance the MI value of negative characteristics and feature’s frequency,supports the concept of concentration degree and dispersion degree.In accordance with the concept of concentration degree and dispersion degree,formulas which embody concentration degree and dispersion degree were constructed and the improved mutual information was implemented based on these.In this paper,the feature selection algorithm was applied based on improved mutual information to a text classifier based on Biomimetic Pattern Recognition and it was compared with several other feature selection methods.The experimental results showed that the improved mutual information feature selection method greatly enhances the performance compared with traditional mutual information feature selection methods and the performance is better than that of information gain.Through the introduction of the concept of concentration degree and dispersion degree,the improved mutual information feature selection method greatly improves the performance of text classification system.     

基于小波变换的多聚焦图像自适应融合

图像融合是对来自同一场景的不同源图像的信息进行互补和合成,从而获得更为准确、更为全面、更为可靠的图像。采用了一种基于小波变换的自适应图像融合方法,首先将配准好的图像进行小波分解,并提取出细节分量和近似分量。其次,针对不同的频率域选择不同的融合规则,对低频系数选取区域均匀度和变化率相结合的融合规则,对高频系数选用区域方向对比度和区域匹配度相结合的自适应融合规则。最后通过小波逆变换得到融合图像。将其它融合算法和文中所提算法进行主观和客观的对比,结果表明,该算法是一种有效可行的图像融合算法。     

Oil monitoring methods based on information theory

To evaluate the wear condition of machines accurately,oil spectrographic entropy,mutual information and ICA analysis methods based on information theory are presented. A full-scale diagnosis utilizing all channels of spectrographic analysis can be obtained. By measuring the complexity and correlativity,the characteristics of wear condition of machines can be shown clearly. The diagnostic quality is improved. The analysis processes of these monitoring methods are given through the explanation of examples. The availability of these methods is validated and further research fields are demonstrated.     

Covert data transmission mechanism based on dynamic label in blockchain

The existing blockchain data covert transmission scheme is faced with the problems of poor concealment, small channel capacity and low efficiency. This paper designs a new model for separating the data embedding mechanism from the transaction filtering mechanism, and proposes a new filtering algorithm based on the address dynamic label. In this scheme,the HMAC algorithm is used to construct the special address label, so that the receiving party can effectively screen the special transaction carrying the concealed information from the massive data in the blockchain ledger, and realize the concealed data transmission in the open blockchain scenario. Simulation results show that the proposed scheme can realize the hidden transmission of data in the real bitcoin network. The concealment and efficiency of the scheme are better than those of the traditional scheme.     

Maximum likelihood registration for systemic error based on statistical linear regression

Generally,there are non-random systemic errors in target detection with the cooperative multi-sensor system.In order to solve this problem,a maximum likelihood registration algorithm based on statistical linear regression (SLR-MLR) is presented.The registration equation for the multi-sensor system is established first by jointly maximizing the likelihood function of the target state and systemic error,on the basis of which the proposed algorithm utilizes a set of diverse regression points to handle the linearization problem of the nonlinear measurement transformation.The regression equation for the target state with respect to unbiased measurement is constructed through statistical linear regression,and then the first two statistical properties of the projected state can be obtained.Moreover,the algorithm uses the likelihood maximization iteration to seek the solution of the registration equation,thus achieving the joint estimation for the systemic error and target state.Simulation results show that the SLR-MLR can achieve the registration of multiple sensors in each observation dimension,and has a higher accuracy and near computational complexity compared with the classical MLR.     

基于互信息的特征子集选择

通过互信息的思想提出一个新的评价函数来评价属性之间的相关性,并结合LV算法进行特征子集选择.结果表明,该方法对分类问题效果明显.     

基于H.264标准的视频混沌密写算法

基于目前最新的视频编码标准H.264,运用混沌加密和混沌密写,提出一种视频混沌密写算法。采用标准视频序列进行了仿真,并分析了算法的性能。结果表明:该算法不仅能快速和有效地进行密写,而且解码后视频序列的视觉效果和峰值信噪比(PSNR)基本上不受密写影响。     

基于双字Hash机制的交通信息分词算法研究

互联网页面文本挖掘是大数据时代快速获取广域实时专题信息的重要技术手段.针对公众出行服务和交通相关应用,互联网页面中所蕴含的交通信息极具价值,但需要借助有效的自然语言分词方法来实现语义理解过程.针对互联网文本页面中蕴含交通信息的表达特点,设计了一种基于双字Hash和List相结合的三层词典数据结构,提出了一种改进的最大匹配分词算法,该算法在切分过程中增加了对关键词汇的词库归属性判断,保存了根据各个词库切分出来的关键词汇的个数与顺序,使其能够更好地满足自然语言语义理解的需求;同时,算法考虑了交通信息词库中词的长度分布,采用分治方法实现对长句或组合句的有效处理.实验表明:所提出的互联网文本页面蕴含交通信息的中文分词算法具有较好的精度和效率,可以满足互联网页面蕴含交通信息的实时分词需求,为交通信息文本的语义理解奠定了技术基础.     

基于MCMC算法贝叶斯网络的学习

对于给定的阈值,通过计算变量之间的互信息,设计了一种构造贝叶斯网络结构的方法。改进了关于图模结构学习中常见的 MCMC 算法。将这种方法构造的贝叶斯网络作为马尔可夫链初始状态的网络结构,利用改进后的 MCMC 算法,构造一个关于贝叶斯网络结构的马尔可夫链。迭代给定次数后,得到关于变量组的贝叶斯网络结构。实验结果表明:改进前和改进后的两种方法得到的贝叶斯网络结构基本一致,网络结构的接受率也相近。     

一种新的红外焦平面阵列疵点补偿算法

提出了一种基于运动图像分析的红外焦平面阵列疵点补偿算法,它利用序列图像帧间的相关性,通过对相邻两帧图像进行运动分析,获取图像中目标的运动参数,再通过运动参数,将当前图像中的疵点对应到前一帧图像相应的位置,并用前一帧相应位置像素的灰度对疵点进行补偿．该算法克服了邻域疵点补偿算法无法保持目标边缘的缺点,其补偿效果优于邻域补偿算法的效果．     

一种基于模糊信息熵的协同过滤推荐方法

针对评分数据的稀疏性制约协同过滤推荐性能的情况,提出一种新的相似性度量方法。首先,定义了用户的模糊信息熵以反映用户评分偏好的不确定程度;其次,利用两两用户的模糊互信息衡量用户之间的相似程度;最后,同时考虑用户之间的模糊互信息和用户的模糊信息熵,并设计一种基于模糊信息熵的相似性度量方法以计算用户之间的相似性。在两个公开数据集上的试验结果表明:基于模糊信息熵的相似性度量方法能够降低数据稀疏性的影响,并能显著提高推荐系统的推荐性能。     

A bidirectional feature selection method based on mutual information and redundancy-synergy coefficient

Featuresubsetselection(FSS)istoselectrele vantfeaturesandcastawayirrelevantandredundantfeaturesfromtheoriginalfeaturesetaccordingtoaFSSmeasure[1].IfafeaturesubsetsatisfiestheFSSmeas ureandhastheminimumsize,itisconsideredapartoftheoptimalfeaturesubset.Comp…