基于DE-EDA-SVM的瓦斯浓度预测建模仿真研究

瓦斯浓度作为衡量煤矿瓦斯危害程度的一个重要指标,为了能够更加准确的预测煤矿瓦斯的浓度,提出一种差分进化-分布估计（DE-EDA）算法优化的支持向量机瓦斯浓度预测新方法。利用无线传感网络系统采集工作环境中的瓦斯浓度数据,并经过降噪处理后作为训练样本。采用DE-EDA算法对SVM模型的惩罚参数C、损失参数ε以及径向基参数γ进行优化,利用优化后的模型进行瓦斯浓度的预测。通过MATLAB软件仿真可以得出,所采用的优化模型能够准确的预测煤矿瓦斯浓度的变化趋势。并与经过粒子群（PSO）算法优化的预测模型相比较。结果表明,经过DE-EDA算法优化的SVM模型具有训练速度更快、预测更准确的特点,为实际煤矿瓦斯浓度的预测和处理提供了更加可靠的理论基础。     

高光谱影像的BDT-SVM地物分类算法与应用

面对海量数据的特征空间高维性及训练样本的有限性,高光谱遥感影像若采用常规统计模式的分类方法难以获得较好的分类结果。因此探讨支持向量机(SVM)分类器的基本原理,针对EO-1Hyperion高光谱影像的分类特点及现有多类SVM算法所存在的训练时间长及分类精度低等问题,引入二叉决策树SVM(BDT-SVM)分类算法,并提出一种新的类间分离度定义方法及相应的客观确定二叉树结构的策略,由此生成改进的BDT-SVM算法。实验结果表明:与其他多类分类方法相比,基于改进的BDT-SVM算法的高光谱影像地物分类效果更好,总体精度达到90.96%,Kappa系数为0.89,该算法还解决了经典SVM多类分类可能存在的不可分区域问题。     

同振式三轴向矢量水听器设计与实现

根据实际工程需要,设计、制作同振式三轴向矢量水听器,其外形为柱体两端带半球帽的圆柱体,声压通道和矢量通道在结构上结合为一体。矢量水听器外形尺寸为Φ54×72 mm、工作频带为20 Hz~2000 Hz。在国防科技工业水声一级计量站驻波管对研制的矢量水听器进行了测试,测试结果表明：声压通道声压灵敏度级为-192.7 dB（0 dB=1 V/mPa）、具有全指向性;矢量通道声压灵敏度级为-185.3 dB（测试频率500 Hz,0 dB=1 V/mPa）、具有余弦指向性。设计、制作的矢量水听器灵敏度高、通道一致性好、使用方便,满足工程应用需求。     

一种改进的融合关联词典的微博倾向性分析方法

大多数研究者对微博倾向性分析过多关注的是情感词、形容词和否定词,忽略了 关联词对其情感倾向的影响。为了提高微博情感倾向性分析的准确率,提出了融合关联词的微博倾向性分析方法,考虑微博文本中形容词、程度副词以及关联词之间的组合关系。 本文充分考虑了关联词的结构特点并在已有词典的基础上构建专门用于微博倾向性分析的微博词典、否定词词典和关联词词典,同时考虑到网络新词对微博倾向性的影响,还构建 了一个全新的网络新词词典。借助支持向量机（Support vector machine,SVM）将微博文本分为负向、正向和中性3 类,通过结合情感词典和SVM的方法提高微博文本倾向性分析的准确率。通过对COASE 2014 数据实验可以表明,本文方法对微博倾向性分析取得了较好的效果。     

一种基于Matrix的QR分解向量化方法

提出一种基于Matrix的Givens旋转的QR分解向量化方法。针对Matrix的体系结构特点,对向量数据访存和计算进行优化,使计算均衡分布到各个向量处理单元;设计双缓冲DMA的数据传输策略,使得内核的计算与DMA数据搬移的时间完全重迭,内核始终处于峰值计算,从而取得最佳的计算效率。实验结果表明,该方法能够取得较高的计算效率和性能加速比。     

Joint underwater target detection and tracking with the Bernoulli filter using an acoustic vector sensor

In this paper, we study the problem of joint underwater target detection and tracking using an acoustic vector sensor (AVS). For this challenging problem, first a realistic frequency domain simulation is set up. The outputs of this simulation generate the two dimensional FRequency–AZimuth (FRAZ) image. On this image, the random finite set (RFS) framework is employed to characterize the target state and sensor measurements. We propose to use the Bernoulli filter, which is the optimal Bayes filter emerged from the RFS framework for randomly on/off switching single dynamic systems. Moreover, to increase the performance of detection and azimuth tracking in low signal-to-noise ratio (SNR) scenarios, a track-before-detect (TBD) measurement model for AVS is proposed to be used with the Bernoulli filter. Sequential Monte Carlo (SMC) implementation is preferred for the Bernoulli filter recursions. Extensive simulation results prove the performance gain obtained by the proposed approach both in estimation accuracy and detection range of the system.     

A decision support system for updating and incorporating climate change impacts into rainfall intensity-duration-frequency curves: Review of the stakeholder involvement process

Stakeholder involvement can serve to increase the quality of decision support systems (DSSs) and increase the perceived legitimacy of DSS outputs. Involving those who are ultimately affected by the outputs of DSSs in system design and development also reflects democratic principles. Importantly, stakeholder involvement can help ensure that the outputs of DSSs are used in decision-making processes. However, DSSs often fail due to poor engagement of stakeholder and end-user communities in the development and design of systems. The stakeholder engagement process applied in the development of the Computerized Tool for the Development of Intensity Duration Frequency Curves under Climate Change described here followed many of the tenants of best practices identified in the literature. While the engagement strategy was generally considered successful, over- and under-representation of some stakeholder groups and long term funding issues were weaknesses in the engagement process.     

Analysis of General Ideal Proportional Navigation Guidance Laws

In this work, the capture region of the general ideal proportional navigation guidance law is analyzed with the following constraints: (i) the target's input acceleration is subject to independent or magnitude saturation, and (ii) the missile's input acceleration is subject to magnitude saturation. In addition to the case of unbounded maneuverability of missile with or without knowing the target acceleration, the necessary and sufficient condition for a bounded maneuvering missile that captures a nonmaneuvering target is also derived. Furthermore, the conditions that result in zero impact velocity are discovered and discussed in some of the cases. The capture regions that cannot be obtained analytically are determined numerically by a powerful classifier, the least square support vector machines. To have the capture region ready for least square support vector machines, all the state variables are transformed into modified polar variables and non‐dimensionalized to reduce the number of independent variables. In order to make the implementation possible in real time, an approximation of the Gaussian radial basis function is adopted to obtain the corresponding nonlinear feature mapping function. Through numerous numerical experiments, the proposed technique is found to be adequate for storing the capture region in an onboard fire control computer.