一种基于用户识别特征模型的中医药健康养老信息推送算法

通过分析传统协同过滤算法的优缺点,设计构建用户识别特征模型,并提出基于用户识别特征模型的协同过滤算法,将经典协同过滤算法与基于用户识别特征模型的协同过滤算法进行比较试验,分析基于用户识别特征模型的协同过滤算法相对于经典协同算法的优势,进而证明其能够在中医药健康养老信息的推送过程中提高效率。     

基于支持向量机的非侵入式负荷印记识别模型

基于非侵入式电力负荷检测与分解技术近年来得到广泛推广.选取14个稳态指标作为负荷特征,建立基于支持向量机(SVM)的非侵入式负荷印记识别模型,利用多分类支持向量机(multi-class SVM)的成对分类算法,对负荷印记进行了识别,随机抽取数据进行测试,结果表明方法能够更准确地识别负荷印记,说明所提出的模型和方法具有较高的有效性和正确性.     

基于多目标进化聚类的信用风险特征识别

在信用风险识别领域,聚类算法常被用于区分不同风险等级的样本并识别风险特征。然而该领域中通常面临高维数据处理问题,导致传统聚类算法存在不适应此类问题的缺陷:易陷入局部最优、受冗余特征干扰、鲁棒性不强等。采用高维信用风险数据,研究上市公司信用风险,建立信用风险特征识别的三目标优化模型,设计基于分解的多目标子空间聚类算法进行求解。通过算法的横向对比实验,展示了所提出的算法在聚类精度和鲁棒性方面的优势,并根据聚类算法的权重分配结果,归纳总结上市公司信用风险评估过程中应重点关注的指标。     

基因识别及其算法研究

针对基因识别问题,基于DNA序列的3周期这一性质,首先给出了DNA序列功率和信噪比的快速算法并讨论了不同物种基因类型的阈值确定方法;在此基础上,建立了基于背景噪声抑制和频谱平滑的SNR频谱预处理模型,经过预处理后的频谱不仅大幅度抑制了背景噪声,同时保留了SNR频谱的模式特征.在编码序列识别上,对经典的EPND预测算法进行了改进,使用改进的EPND算法对经过预处理后频谱进行基因识别,实验结果显示这种基因识别模型具有优异的基因识别性能,比传统直接使用基于滑动窗口DFT的EPND识别算法在敏感度、特异性等评价指标上提高了2%-12%左右.     

基于改进Gentle-Adaboost算法的疲劳驾驶中人脸检测研究

疲劳驾驶检测中人脸检测的算法有很多,其中相对成熟的是Gentle-Adaboost,此算法虽然识别率较高,但是识别时间较长,研究在识别过程中使用均值哈希算法进行算法优化,并引入缓存的概念来缩短识别时间.首先,根据Gentle-Adaboost算法和Haar特征来提取基本特征点,然后通过对比前后两帧图像的哈希指纹来减少人脸检测次数,最后通过缓存数据库来存储相似图片的哈希指纹,运行一段时间后仅需通过对比哈希指纹就能精准的找到人脸区域.通过实验,可以证明改进后的平均识别时间可减少原时间的80%.     

数据空间结构性及在KNN算法中的应用

数据空间结构性是多维数据客观存在的本征特征,是数据挖掘的重要内容.通过统计学的方法,分析多维数据空间属性变量和类型变量之间的结构特征,可以准确刻画数据在多维变量空间的相关性及其各向异性.数据空间结构特征可以用于机器学习算法的改进和提升,以提高模式识别的效果.融合了数据空间结构特征的KNN算法在稳定性和识别精度上均优于传统算法.通过在苏里格气田苏东41-33区块复杂碳酸盐岩的岩性识别中的应用表明,与传统KNN算法相比,数据空间结构的引入能提高识别准确率12.35%,并显示出算法的灵活性和适用性.多维数据空间结构的研究对机器学习算法的泛化能力和迁移性的提升等方面具有促进作用.     

基于直觉模糊熵权和CC-OWA算子的雷达目标识别模型

为更完整的描述和表达雷达目标类型识别中的目标特征和目标类型之间的关系复杂性和知识缺乏性,通过直觉模糊关系描述,进而将目标识别特征信息转化为直觉模糊集信息.分析了基于直觉模糊集理论的雷达目标类型识别知识建模,揭示了直觉模糊信息的价值可以通过直觉模糊熵刻画,进而提出应用直觉模糊集的熵构造特征直觉模糊信息的权重(直觉模糊熵权),充分利用了目标类型识别知识中隐含的权重信息,并结合CC-OWA算子建立雷达目标类型识别模型与识别步骤,利用一个雷达目标识别实例说明了模型的有效性.     

线性双层规划的改进PCP全局求解算法

本文采用K-T条件将线性双层规划模型改写为单层规划后,将参数引入上层目标函数,构造了含参线性互补问题(PLCP)并给出它的一些性质。进而通过改进Lemke算法的进基规则,在保持互补旋转算法原有优势的基础上,引入充分小正数ε,设计了改进参数互补旋转(PCP)算法求取全局最优解,最后通过两个算例说明了其有效性。     

多模型加权融合机制的石墨纯度识别

针对人工识别的效率低及单一卷积神经网络提取特征的遗漏问题,提出了多模型加权融合机制的石墨纯度识别算法.在自建小样本数据集上,进行离线扩充和在线增强,提高模型的泛化能力,减少深层CNN的过拟合问题;结合迁移学习,利用优化的AlexNet和ResNet50构建双通道卷积神经网络,提取石墨图像的深层次特征,并将两者的特征进行...     

ABC-PSO算法优化混合核SVM参数及应用

针对混合核支持向量机(SVM)中的可调参数一般是根据经验或人工随机调试得到,不能确保参数最优的局限性,提出用粒子群和人工蜂群的并行混合优化(ABC-PSO)算法来优化混合核SVM参数,找出满足条件的最优参数组合.将该SVM模型应用到语音识别中,通过对三个不同语种的语音数据库的实验仿真,验证了混合算法优化SVM参数所得的优化SVM模型比PSO算法优化SVM所得的模型,具有良好的泛化能力和语音识别能力.     

Recognition of triangles by covariogram

The paper considers the problem of recognition of triangles by orientation-dependent chord length distribution. The following results are obtained: 1. The explicit form of the covariogram and orientation-dependent chord length distribution function for a triangle. 2. The explicit form for the chord length distribution function for a triangle. 3. The length of the maximal chord of a triangle is continuous function on direction u ∈ S ¹ (S ¹ is the space of all directions in the plane). 4. If we have orientation-dependent chord length distribution function for an everywhere dense set of S ¹, then we can uniquely recognize the triangle with respect to reflections and translations. 5. For any finite subset A ? S ¹, there are two non-congruent triangles with the same values of orientationdependent chord length distribution functions on A.     

Chord Diagrams, d-Diagrams, and Knots

A method of representing knots, links, and singular knots and links by words in a finite alphabet and so-called d-diagrams is given. A representation of the chord diagram algebra by words in a finite alphabet is considered. This method, unlike coding by Gauss diagrams, allows one to consider knots and links simultaneously. An algorithm for recognition of diagrams of classical knots in terms of d-diagrams is given. Bibliography: 9 titles.     

R3中双弦幂积分

本文研究R3中凸体的双弦幂积分,利用线偶的运动密度,得到了一些性质及球体上的部分双弦幂积分值,并由此解决了R3中一几何概率问题.     

Reliability analysis and optimization of weighted voting systems with continuous states input

Weighted voting systems are widely used in many practical fields such as target detection, human organization, pattern recognition, etc. In this paper, a new model for weighted voting systems with continuous state inputs is formulated. We derive the analytical expression for the reliability of the entire system under certain distribution assumptions. A more general Monte Carlo algorithm is also given to numerically analyze the model and evaluate the reliability. This paper further proposes a reliability optimization problem of weighted voting systems under cost constraints. A genetic algorithm is introduced and applied as the optimization technique for the model formulated. A numerical example is then presented to illustrate the ideas.     

Covariogram of a parallelogram

In this paper we obtain the explicit forms of the covariogram and the orientation-dependent chord length distribution function for any parallelogram. The explicit form of the chord length distribution function for a parallelogram is also obtained.     

Blind source separation with nonlinear autocorrelation and non-Gaussianity

Blind source separation (BSS) is a problem that is often encountered in many applications, such as biomedical signal processing and analysis, speech and image processing, wireless telecommunication systems, data mining, sonar, radar enhancement, etc. One often solves the BSS problem by using the statistical properties of original sources, e.g., non-Gaussianity or time-structure information. Nevertheless, real-life mixtures are likely to contain both non-Gaussianity and time-structure information sources, rendering the algorithms using only one statistical property fail. In this paper, we address the BSS problem when source signals have non-Gaussianity and temporal structure with nonlinear autocorrelation. Based on the two statistical characteristics of sources, we develop an objective function. Maximizing the objective function, we propose a gradient ascent source separation algorithm. Furthermore, We give some mathematical properties for the algorithm. Computer simulations for sources with square temporal autocorrelation and non-Gaussianity illustrate the efficiency of the proposed approach.     

Total chromatic number of unichord-free graphs

A unichord is an edge that is the unique chord of a cycle in a graph. The class C of unichord-free graphs — that is, graphs that do not contain, as an induced subgraph, a cycle with a unique chord — was recently studied by Trotignon and Vuškovi? (2010) [24], who proved strong structure results for these graphs and used these results to solve the recognition and vertex-colouring problems. Machado et al. (2010) [18] determined the complexity of the edge-colouring problem in the class C and in the subclass C^′ obtained from C by forbidding squares. In the present work, we prove that the total-colouring problem is NP-complete when restricted to graphs in C. For the subclass C^′, we establish the validity of the Total Colouring Conjecture by proving that every non-complete {square, unichord}-free graph of maximum degree at least 4 is Type 1.     

Arc Length Estimation and the Convergence of Polynomial Curve Interpolation

When fitting parametric polynomial curves to sequences of points or derivatives we have to choose suitable parameter values at the interpolation points. This paper investigates the effect of the parameterization on the approximation order of the interpolation. We show that chord length parameter values yield full approximation order when the polynomial degree is at most three. We obtain full approximation order for arbitrary degree by developing an algorithm which generates more and more accurate approximations to arc length: the lengths of the segments of an interpolant of one degree provide parameter intervals for interpolants of degree two higher. The algorithm can also be used to estimate the length of a curve and its arc-length derivatives. AMS subject classification (2000) 65D05, 65D10