Quantification of secondary structure prediction improvement using multiple alignments

The use of multiple sequence alignments for secondary structurepredictions is analysed. Seven different protein families, containingonly sequences of known structure, were considered to providea range of alignment and prediction conditions. Using alignmentsobtained by spatial superposition of main chain atoms in knowntertiary protein structures allowed a mean of 8% in secondarystructure prediction accuracy, when compared to those obtainedfrom the individual sequences. Substitution of these alignmentsby those determined directly from an automated sequence alignmentalgorithm showed variations in the prediction accuracy whichcorrelated with the quality of the multiple alignments and distanceof the primary sequence. Secondary structure predictions canbe reliably improved using alignments from an automatic alignmentprocedure with a mean increase of 6.87percnt;, giving an overallprediction accuracy of 68.5%, if there is a minimum of 25% sequenceidentity between all sequences in a family.     

Trickle Flow Multiplicity: The Influence of the Prewetting Procedure on Flow Hysteresis

The existence of multiple hydrodynamic studies (MHS) in trickle flow is a well-known phenomenon. It is also known that different prewetting procedures result in major differences in MHS when the hydrodynamic variables pressure drop, liquid holdup and gas–liquid mass transfer are considered. Given a certain prewetting procedure one still has the option to perform flow hysteresis cycles to achieve an even wider variety of MHS. Although numerous studies have been performed on trickle flow hysteresis, none have attempted to decouple the hysteresis behaviour from the prewetting procedure followed. Accordingly there are numerous hysteresis possibilities that have not been investigated. In this work a single liquid and gas cycle were performed for four distinct prewetting procedures described here as a dry bed, a Levec type prewetted bed, Kan prewetted bed (achieved by increasing either the liquid or the gas flow rate until the pulsing flow regime is reached) and a Super prewetted bed. Pressure drop, liquid holdup and gas–liquid mass transfer are the hydrodynamic parameters studied to quantify the various MHS. It is shown that the shape and extent of the hysteresis cycle are strongly dependant on the prewetting procedure. In terms of flow structure, similar hysteresis trends on the Kan Liquid and Super prewetting modes indicate that these modes are hydrodynamically similar. The additional measurement of the hysteresis behaviour of gas–liquid mass transfer proofs that neither holdup nor pressure drop can be used as an indicator of the distribution uniformity.     

域自适应学习研究进展

传统的机器学习假设测试样本和训练样本来自同一概率分布. 但当前很多学习场景下训练样本和测试样本可能来自不同的概率分布. 域自 适应学习能够有效地解决训练样本和测试样本概率分布不一致的学习问题,作为 机器学习新出现的研究领域在近几年受到了广泛的关注. 鉴于域自适应学习技术 的重要性,综述了域自适应学习的研究进展. 首先概述了域自适应学习的基本问 题,并总结了近几年出现的重要的域自适应学习方法. 接着介绍了近几年提出的 较为经典的域自适应学习理论和当下域自适应学习的热门研究方向,包括样例加 权域自适应学习、特征表示域自适应学习、参数和特征分解域自适应学习和多 源域自适应学习. 然后对域自适应学习进行了相关的理论分析,讨论了高效的度 量判据,并给出了相应的误差界. 接着对当前域自适应学习在算法、模型结构和 实际应用这三个方面的研究新进展进行了综述. 最后分别探讨了域自适应学习在 特征变换和假设、训练优化、模型和数据表示、NLP 研究中存在的问题这四个方面 的有待进一步解决的问题.     

多点多板多分型面注塑模具的简化设计

应用锁扣装置、组合拉杆结构，减少导柱数量，简化了多点多板多分型面注塑模具的设计。     

面向ZigBee的视频多描述编码及其传输性能仿真

针对视频码流在ZigBee等不可靠低速无线网络传输时易发生编码信息包丢失而导致接收端无法正确解码这一问题,提出一种基于3D-SPIHT视频多描述编码方案。与传统多描述编码如MDSQ相比,该方案创建的各描述质量平衡,并能通过去除视频帧间冗余来提高压缩比;与国际标准（H.264等）相比,该方案具有算法简单、能够在任意码流截断处解码、码率可连续分级和无明显块效应等优点。实验表明,该编码方案能够较好地应用于ZigBee多信道网络,并在较大压缩比的情况下仍能获取质量较好的重建图像,具有较高的应用价值。     

基于动态射频指纹的调频定位方法

针对调频(FM)广播信号动态变化对定位性能产生较大影响的问题,提出了动态射频指纹的FM定位方法。该方法采用多元线性回归和神经网络方法,根据离线阶段各个参考点和部分较准点的调频广播信号强度的关系,使用在线阶段较准点的信号强度实时估计在线阶段参考点的信号强度。通过这两种方法建立具有自适应能力的动态射频地图,并使用贝叶斯估计方法对目标进行定位。实验结果表明,相对静态射频指纹模型：采用多元线性回归的射频指纹动态映射模型定位误差平均减少9.1%,采用神经网络的射频指纹动态映射模型定位误差平均减少36.3%,有效抑制了射频信号动态变化时变性对定位性能的影响。     

基于乘法偏好关系的群一致性偏差熵多属性群决策方法研究

在多属性群决策中, 针对每一个属性下决策者都有一个关于决策方案的乘法偏好关系的决策问题, 提出一种基于乘法偏好关系的群一致性偏差熵多属性群决策方法. 此方法考虑到不同属性下的决策者具有不同的权重, 并通过迭代运算可以达到群一致性水平, 从而得出最终的不同属性下决策者的权重; 同时, 可以利用偏差熵模型来求解属性权重, 利用这两个权重最终获得一个综合各方意见的群一致性乘法偏好关系. 最后通过算例分析验证了所提出方法的有效性.     

混合估计多模粒子滤波的机动弱小目标检测前跟踪算法

针对低信噪比条件下的机动小目标实时检测与跟踪困难的问题, 提出一种基于混合估计多模粒子滤波的检测前跟踪改进算法. 首先根据前一时刻所采用的模型状态及其转移概率等先验信息实现当前时刻的模型采样; 然后在充分考虑当前量测下实现当前的粒子预测, 采用一种序贯重要性平滑重采样策略, 在不增加计算量的前提下, 改善了粒子多样性衰退的问题; 最后通过新的粒子集完成对模型和状态的合理估计与目标检测. 仿真结果验证了该方法的检测与跟踪性能优于传统的多模粒子滤波方法.

     

基于四元数和增广矩阵束的MIMO雷达角度估计算法

通过研究多输入多输出（Multiple input and multiple out, MIMO）雷达的角度估计算法,基于收发共址的十字阵MIMO雷达系统,将四元数理论应用到 MIMO雷达角度中,提出了一种新的参数估计算法。通过构造四元数模型,结合增广矩阵束（ Matrix enhancement and matrix pencil, MEMP）方法构造增广矩阵,并证明该矩阵的秩 等于目标总数,且不受目标相干性的影响,结合ESPRIT算法实现了对MIMO相干目标的角度估 计。算法无需谱峰搜索,能够实现参数的自动配对,同时降低了运算复杂度。仿真实验进一 步验证了算法的有效性。