基于K-means算法的建筑群震害分析模型缩减方法

基于建筑群模型和弹塑性时程分析的精细化城市震害模拟技术能够为防震减灾及应急救援决策提供必要的依据和参考。为了减小城市建筑群震害模拟的计算量和计算时间,本文提出一种基于聚类算法的建筑群模型缩减方法。该方法采用K-means聚类算法,首先基于建筑结构属性向量对建筑群进行聚类,将相似的建筑结构聚为一组;然后从每组选取一个代表建筑组成建筑群缩减模型,通过减少需要分析的建筑结构数量来减少建筑群震害模拟的计算量。本文对传统的K-means算法进行改进,通过设定组内建筑结构的差异上限自动调整聚类分组数量;提出将具体地震动作用下结构地震损伤指数作为结构属性向量进行聚类,并通过算例对比分别采用两种缩减模型,即基于损伤指数聚类的缩减模型与基于结构力学模型参数聚类的缩减模型,计算结构损伤状态准确程度。对比结果表明:在聚类分组数量相同的情况下,基于损伤指数的分组明显优于基于模型参数的分组,采用模型缩减方法能够在保证足够计算精度前提下显著减少建筑群震害模拟计算量和计算时间。

Model update mechanism for mean-shift tracking

In order to solve the model update problem in mean-shift based tracker, a novel mechanism is proposed. Kalman filter is employed to update object model by filtering object kernel-histogram using previous model and current candidate. A self-tuning method is used for adaptively adjust all the parameters of the filters under the analysis of the filtering residuals. In addition, hypothesis testing servers as the criterion for determining whether to accept filtering result. Therefore, the tracker has the ability to handle occlusion so as to avoid over-update. The experimental results show that our method can not only keep up with the object appearance and scale changes but also be robust to occlusion.

Portfolio Choice under the Mean-Variance Model with Parameter Uncertainty

Assuming the investor is uncertainty-aversion,the multiprior approach is applied to studying the problem of portfolio choice under the uncertainty about the expected return of risky asset based on the mean-variance model. By introducing a set of constraint constants to measure uncertainty degree of the estimated expected return,it built the max-min model of multi-prior portfolio,and utilized the Lagrange method to obtain the closed-form solution of the model,which was compared with the mean-variance model and the minimum-variance model; then,an empirical study was done based on the monthly returns over the period June 2011 to May 2014 of eight kinds of stocks in Shanghai Exchange 50 Index. Results showed,the weight of multi-prior portfolio was a weighted average of the weight of mean-variance portfolio and that of minimumvariance portfolio; the steady of multi-prior portfolio was strengthened compared with the mean-variance portfolio; the performance of multi-prior portfolio was greater than that of minimum-variance portfolio. The study demonstrates that the investor can improve the steady of multi-prior portfolio as well as its performance for some appropriate constraint constants.

Calculation of Interaction Parameters from Immiscible Phase Diagram of Alkali Metal or Alkali Earth Metal-Halide System by Means of Subregular Solution Model

In this paper, the interaction parameters in the subregular solution model, λ1 and λ2, are regarded as a linear function of temperature, T. Therefore, the molar excess Gibbs energy of A-B binary system may be reexpressed as follows:Gm^E=xAxB[(λ11+λ12T)+(λ21+λ22T)xB]The calculation of the model parameters, λ11, λ12, λ21and λ22, was carried out numerically from the phase diagrams for 11 alkali metal-alkali halide or alkali earth metal-halide systems. In addition, artificial neural network trained by known data has been used to predict the values of these model parameters. The predicted results are in good agreement with the .calculated ones. The applicability of the subregular solution model to the alkali metal-alkali halide or alkali earth metal-halide systems were tested by comparing the available experimental composition along the boundary of miscibility gap with the calculated ones which were obtained by using genetic algorithm. The good agreement between the calculated and experimental results across the entire liquidus is valid evidence in support of the model.

A New Mean Reversion Model of Close-End Fund

On the basis of fractal theory, one of the nonlinear theories, this paper studies the validity of Chinese fund market fractal time sequence through Hurst exponent, calculates the H value and proposes a new close-end fund mean reversion model. Meanwhile, this paper validates the mean reversion time sequence for consecutive 54 week data of fund market. The result indicates that this model can effectively prove that Chinese close-end fund market follows the biased random walk. The research also proves that the fund discount does have mean reversion tendency and averagely the fund with high discount has a higher excess yield than that of the fund with low discount. The mean excess yield and the ratio between discount rate deviation and standard deviation demonstrate a descending relationship. The optimum investment period based on ＂mean reversion＂ is one month. Consequently this model provides a new arbitrage method through the discount of close-end fund.

Joint Variable Selection of Mean-Covariance Model for Longitudinal Data

In this paper we reparameterize covariance structures in longitudinal data analysis through the modified Cholesky decomposition of itself. Based on this modified Cholesky decomposition, the within-subject covariance matrix is decomposed into a unit lower triangular matrix involving moving average coefficients and a diagonal matrix involving innovation variances, which are modeled as linear functions of covariates. Then, we propose a penalized maximum likelihood method for variable selection in joint mean and covariance models based on this decomposition. Under certain regularity conditions, we establish the consistency and asymptotic normality of the penalized maximum likelihood estimators of parameters in the models. Simulation studies are undertaken to assess the finite sample performance of the proposed variable selection procedure.

Monthly Mean Temperature Prediction Based on a Multi－level Mapping Model of Neural Network BP Type

In terms of 34-year monthly mean temperature series in 1946-1979,the multi-level maPPing model of neural netWork BP type was applied to calculate the system's fractual dimension Do=2'8,leading tO a three-level model of this type with ixj=3x2,k=l,and the 1980 monthly mean temperture predichon on a long-t6rm basis were prepared by steadily modifying the weighting coefficient,making for the correlation coefficient of 97% with the measurements.Furthermore,the weighhng parameter was modified for each month of 1980 by means of observations,therefore constrcuhng monthly mean temperature forecasts from January to December of the year,reaching the correlation of 99.9% with the measurements.Likewise,the resulting 1981 monthly predictions on a long-range basis with 1946-1980 corresponding records yielded the correlahon of 98% and the month-tO month forecasts of 99.4%.

Exact analytical solutions to the mean-field model depicting microcavity containing semiconductor quantum wells

By using a two-mode mean-field approximation, we study the dynamics of the microcavities containing semiconductor quantum wells. The exact analytical solutions are obtained in this study. Based on these solutions, we show that the emission from the microcavity manifests periodic oscillation behaviour and the oscillation can be suppressed under a certain condition.

基于K-means算法的跨国零售商客户细分研究

随着经济全球化及大数据技术的蓬勃发展,跨国零售商之间的竞争日益激烈,根据客户特征进行客户细分,协助客户进行个性化的服务体验,有利于跨国零售商实现精准营销和高效的客户关系管理。为了提高客户细分的精度,本文提出一种基于RFM模型的K-means聚类算法,使用簇内误方差(SSE)和轮廓系数(Silhouette Coefficient)计算聚类个数,优化K值选取。本文选取一家跨国零售商的数据进行实证检验,对细分后的结果进行特征分析,将客户划分为核心型客户、维护型客户和风险型客户三种类别,并为不同客户群体提供差异化营销策略,仅供参考。

A Fractional Micro-Macro Model for Crowds of Pedestrians Based on Fractional Mean Field Games

Modeling a crowd of pedestrians has been considered in this paper from different aspects. Based on fractional microscopic model that may be much more close to reality, a fractional macroscopic model has been proposed using conservation law of mass. Then in order to characterize the competitive and cooperative interactions among pedestrians, fractional mean field games are utilized in the modeling problem when the number of pedestrians goes to infinity and fractional dynamic model composed of fractional backward and fractional forward equations are constructed in macro scale. Fractional micromacro model for crowds of pedestrians are obtained in the end. Simulation results are also included to illustrate the proposed fractional microscopic model and fractional macroscopic model, respectively. © 2014 Chinese Association of Automation.

基于CoSENT和改进K-means的冒犯性评论文本主题识别

为快速识别冒犯性评论文本中的用户热点主题,解决传统主题模型在处理评论文本时语义描述不充分、上下文信息丢失和主题连贯性不强,以及K-means聚类算法对K值和初始中心点敏感的问题。使用CoSENT(cosine sentence)模型获取包含冒犯性语言的评论文本的句子级向量特征,对通过统一流形逼近与投影算法即UMAP(uniform manifold approximation and projection)模型降维后的向量矩阵使用基于Canopy+的改进K-means算法进行类簇划分,用(class term frequency-inverse document frequency,c-TF-IDF)识别各主题簇的主题特征,进行主题建模。通过对比冒犯性评论文本数据集以及普通评论数据集的实验验证了方法有效性。结果表明本文方法能够得到更好的主题一致性。

A Mean-variance Problem in the Constant Elasticity of Variance(CEV) Model

In this paper, we focus on a constant elasticity of variance （CEV） modeland want to find its optimal strategies for a mean-variance problem under two constrainedcontrols： reinsurance/new business and investment （no-shorting）. First, aLagrange multiplier is introduced to simplify the mean-variance problem and thecorresponding Hamilton-Jacobi-Bellman （HJB） equation is established. Via a powertransformation technique and variable change method, the optimal strategies withthe Lagrange multiplier are obtained. Final, based on the Lagrange duality theorem,the optimal strategies and optimal value for the original problem （i.e., the efficientstrategies and efficient frontier） are derived explicitly.

Performance of CMIP6 models in simulating the dynamic sea level:Mean and interannual variance

本研究采用卫星测高数据与第六次国际耦合模式比较计划(CMIP6)海平面动力进行对比,重点针对40S-40N地区的动力海平面(DSL),评估了模式对其平均态与年际变率的综合模拟能力,结果表明,对于DSL平均态的模拟,模式与观测结果非常吻合,模式之间的差异较小.其中,副热带北大西洋是模拟偏差和模式间差异较为显著的区域,对于DSL年际变率的模拟,模式之间保持较高的一致性,但是,模式与观测结果存在明显差异,模式普遍低估了DSL的年际方差;其中,误差大值区域出现在副热带西边界流附近,模式分辨率会影响CMIP6对中小尺度海洋过程的重现能力,这可能是导致CMIP6历史模拟出现误差的原因之一.

电网需求侧资源动态分布式k-means聚类算法

为有效聚合电网需求侧资源,合理、高效利用电网资源,提出基于分布式k-means的电网需求侧资源动态聚类算法。通过基于置信半径的分布式k-means算法聚类采集到的电网需求侧资源数据,在模糊C均值进化神经网络中,以聚类得到的电网需求侧资源数据为输入向量,输出电网需求侧资源场景,依据场景存在概率,以电网侧资源日均峰谷差最小、DG消纳程度最高与日均负荷波动率最小为目标函数,以电网需求侧资源曲线波动率与负荷互补为约束条件,构建电网需求侧资源多场景聚类模型,经动态改变惯性因子(DCW)粒子群算法求解模型后,实现电网需求侧资源多场景聚类。试验结果表明:该方法可实现电网需求侧资源动态聚类,应用该方法聚类不同场景电网需求侧资源时的日负荷率较低,聚类效果较好,可满足实际电力需求侧资源动态聚类工作的需要。

改进的RFM模型和K-means算法在会员分类中的应用研究

针对传统RFM模型用于会员分类会产生失真的问题,对RFM模型提出了改进,增加了客户关系长度和客户购买周期两个参数。同时针对传统的K-means算法存在的问题,提出了一种基于样本对象特征方差加权与中心初始化的K-means算法。利用改进的RFM模型对会员进行分类,可以有效地提高分类效率。

Study of the Magnetocaloric Effect in La0.5Sm0.2Sr0.3Mn1-xFexO3 (x = 0 and 0.05) Manganites with the Mean-Field Theory

In this paper, the magnetocaloric in La0.5Sm0.2Sr0.3Mn1-xFexO3 compounds with x = 0 (LSSMO) and x = 0.05 (LSSMFO) were simulated using mean field model theory. A strong consistency was observed between the theoretical and experimental curves of magnetizations and magnetic entropy changes, −ΔSM(T). Based on the mean-field generated −ΔSM(T), the substantial Temperature-averaged Entropy Change (TEC) values reinforce the appropriateness of these materials for use in magnetic refrigeration technology within TEC (10) values of 1 and 0.57 J∙kg−1∙K−1under 1 T applied magnetic field.

改进K-means聚类的噪声污染监测网络模型的设计与研究

为有效优化城市噪声监测网络,最小化监测点数量,提出了一种改进K-means聚类的噪声污染监测网络模型。通过计算轮廓系数以改进K值,从而改进K-means聚类算法,建立K-means聚类模型,并根据聚类结果的类质心位置确定监测点位置坐标,得到优化的监测网络布局。结果表明,该模型在保证监测结果准确性和全面性的同时,有效减少了监测点的数量,从而降低了监测成本。

基于狄利克雷多项式过程模型与K-means结合的菌群分析

群体分型是一种有助于更好的理解人类身心健康等复杂生物学问题的有效方法,聚类是一种为了对样本分组来降低复杂性的定义肠型的方法,而传统K-means聚类算法的K值选取无法确定,本文在传统K-means聚类算法的基础上进行了改进,并公开数据集上进行了验证,实验表明改进算法能够解决K值选取无法确定的问题,且聚类结果的稳定性、准确性和聚类质量都得到显著提高。将改进后的模型运用于肠道菌群OTUs数据,发现不仅能够有效地区分2-型糖尿病患者样本间的相似性,而且能鉴定出影响菌群结构异质性最大的OTUs菌,为临床解决2-型糖尿病问题提供了一种新的思路。