Mutual Information and Relative Entropy of Sequential Effect Algebras

In this paper,we introduce and investigate the mutual information and relative entropy on the sequentialeffect algebra,we also give a comparison of these mutual information and relative entropy with the classical ones by thevenn diagrams.Finally,a nice example shows that the entropies of sequential effect algebra depend extremely on theorder of its sequential product.

Plotting the Sequential Rock Remote Sensing Information by Optimization Dichotomy Method

The sequential rock remote sensing information is a group of rocks that are correlative in space or in space and time. For the sake of plottiug them, someone had brought forward the optimization segn.entotion metkod. We have ased this method to plot the sequential rock remote sensing information at tbe remote sensing hyperspetral test field of Daqing mountain, Inner Mongolia Autonomous Region, China, and found some disadvantages of this method. Therefore, we put forward the optimization dichotomy to plot them, and get better results. Finally we make a conclusion.

Hierarchical Visualized Multi-level Information Fusion for Big Data of Digital Image

At present,the process of digital image information fusion has the problems of low data cleaning unaccuracy and more repeated data omission,resulting in the unideal information fusion.In this regard,a visualized multicomponent information fusion method for big data based on radar map is proposed in this paper.The data model of perceptual digital image is constructed by using the linear regression analysis method.The ID tag of the collected image data as Transactin Identification(TID)is compared.If the TID of two data is the same,the repeated data detection is carried out.After the test,the data set is processed many times in accordance with the method process to improve the precision of data cleaning and reduce the omission.Based on the radar images,hierarchical visualization of processed multi-level information fusion is realized.The experiments show that the method can clean the redundant data accurately and achieve the efficient fusion of multi-level information of big data in the digital image.

Modified sequential importance resampling filter

In order to deal with the particle degeneracy and impov- erishment problems existed in particle filters, a modified sequential importance resampling （MSIR） filter is proposed. In this filter, the resampling is translated into an evolutional process just like the biological evolution. A particle generator is constructed, which introduces the current measurement information （CMI） into the resampled particles. In the evolution, new particles are first pro- duced through the particle generator, each of which is essentially an unbiased estimation of the current true state. Then, new and old particles are recombined for the sake of raising the diversity among the particles. Finally, those particles who have low quality are eliminated. Through the evolution, all the particles retained are regarded as the optimal ones, and these particles are utilized to update the current state. By using the proposed resampling approach, not only the CMI is incorporated into each resampled particle, but also the particle degeneracy and the loss of diver- sity among the particles are mitigated, resulting in the improved estimation accuracy. Simulation results show the superiorities of the proposed filter over the standard sequential importance re- sampling （SIR） filter, auxiliary particle filter and unscented Kalman particle filter.

Power Allocation for Sensing-Based Spectrum Sharing Cognitive Radio System with Primary Quantized Side Information

Spectrum access approach and power allocation scheme are important techniques in cognitive radio(CR) system,which not only affect communication performance of CR user(secondary user,SU) but also play decisive role for protection of primary user(PU).In this study,we propose a power allocation scheme for SU based on the status sensing of PU in a single-input single-output(SISO) CR network.Instead of the conventional binary primary transmit power strategy,namely the sensed PU has only present or absent status,we consider a more practical scenario when PU transmits with multiple levels of power and quantized side information known by SU in advance as a primary quantized codebook.The secondary power allocation scheme to maximize the average throughput under the rate loss constraint(RLC) of PU is parameterized by the sensing results for PU,the primary quantized codebook and the channel state information(CSI) of SU.Furthermore,Differential Evolution(DE) algorithm is used to solve this non-convex power allocation problem.Simulation results show the performance and effectiveness of our proposed scheme under more practical communication conditions.

Time sequential influence maximization algorithm based on neighbor node influence

In view of the forwarding microblogging,secondhand smoke,happiness,and many other phenomena in real life,the spread characteristic of the secondary neighbor nodes in this kind of phenomenon and network scheduling is extracted,and sequence influence maximization problem based on the influence of neighbor nodes is proposed in this paper.That is,in the time sequential social network,the propagation characteristics of the second-level neighbor nodes are considered emphatically,and k nodes are found to maximize the information propagation.Firstly,the propagation probability between nodes is calculated by the improved degree estimation algorithm.Secondly,the weighted cascade model(WCM) based on static social network is not suitable for temporal social network.Therefore,an improved weighted cascade model(IWCM) is proposed,and a second-level neighbors time sequential maximizing influence algorithm(STIM) is put forward based on node degree.It combines the consideration of neighbor nodes and the problem of overlap of influence scope between nodes,and makes it chronological.Finally,the experiment verifies that STIM algorithm has stronger practicability,superiority in influence range and running time compared with similar algorithms,and is able to solve the problem of maximizing the timing influence based on the influence of neighbor nodes.

融合项目特征级信息的稀疏兴趣网络序列推荐

在以往提取多兴趣嵌入的序列推荐模型中仅能通过聚类的方法发现少量兴趣概念,忽视项目交互序列中特征级信息对最终推荐结果的影响。针对此问题,对传统的多兴趣序列推荐模型进行改进,提出一种融合项目特征级信息的稀疏兴趣网络序列推荐模型。实验结果表明,相比其它模型,该模型可以更好捕捉用户的多样化偏好并缓解冷启动问题。在给定数据集上,该模型比传统的序列推荐模型在命中率上平均提高了6.4%,归一化折损累计增益平均提高了8.7%。

多传感器非线性系统序贯观测融合二阶扩展卡尔曼滤波器

对多传感器非线性系统提出了一种序贯观测融合二阶扩展卡尔曼滤波(Second-order Extended Kalman Filter,SOEKF)算法。该算法的中心思想是根据传感器观测数据到达融合中心的先后次序依次进行处理。研究表明,提出的序贯观测融合SOEKF算法比集中式观测融合EKF算法具有更高的精度;与集中式观测融合SOEKF算法精度相当,且具有更低的计算复杂度。目标跟踪系统的仿真验证了算法的有效性。

基于多空间属性信息融合的序列推荐

序列推荐旨在从用户的历史行为中建模用户不断变化的兴趣,从而做出与用户兴趣相关的推荐。近年来,物品属性信息被证明可以提升序列推荐的性能,很多工作基于属性信息融合去提升序列推荐的性能,都取得了成效但仍存在一定的不足。首先,它们没有显式地建模出用户对物品属性的偏好或者只建模了一个属性偏好向量,无法充分表达用户的偏好。其次,它们的物品属性信息融合过程未考虑用户个性化信息的影响。因此,针对上述不足,提出了基于多空间属性信息融合的序列推荐(MAIF-SR)。文中提出了多空间属性信息融合框架,在不同的属性空间下融合属性序列并建模出用户对不同属性的偏好,用多维兴趣充分表达用户的偏好;设计了个性化属性注意力机制,在融合信息的过程中引入用户个性化信息,增强融合信息的个性化效果。在两个公开数据集以及一个工业私有数据集上进行实验,结果表明,MAIF-SR优于用于对比的基于属性信息融合的序列推荐。

基于三角模糊数序贯博弈的电网安全研究

智能电网系统面临着复杂性、多样性的网络安全风险。针对电网工艺漏洞攻防双方在博弈分析中无法对信息属性损失作出准确判断的问题,提出了一种三角模糊数序贯博弈模型。以序贯博弈树方法进行智能电网攻防博弈,构建出攻击者和防御者的最优策略,实现对智能电网的工控安全防护。通过对智能电网攻击者和防御者构建序贯博弈模型,得到攻防博弈效用。提出运用三角模糊数和概率分布向量的方法来构建非模糊化损失,得到攻防双方清晰化的系统损失值。对三角模糊数序贯博弈算法和贝叶斯序贯算法进行比较,得到智能电网均衡路径中的攻击者和防御者的最优策略,以保障智能电网网络安全运行和防护。该研究为智能电网以及典型工控系统网络安全博弈提供参考。

Design of Irregular QC-LDPC Code Based Multi-Level Coded Modulation Scheme for High Speed Optical Communication Systems

In this paper, we focus on the design of irregular QC-LDPC code based multi-level coded modulation(MLCM) scheme by jointly optimizing the component code rate and the degree distribution of the irregular QC-LDPC component code. Firstly, the sub-channel capacities of MLCM systems is analyzed and discussed, based on which the optimal component code rate can be obtained. Secondly, an extrinsic information transfer chart based two-stage searching algorithm is proposed to find the good irregular QC-LDPC code ensembles with optimal component code rates for their corresponding sub-channels. Finally, by constructing the irregular QC-LDPC component codes from the designed ensembles with the aim of possibly enlarging the girth and reducing the number of the shortest cycles, the designed irregular QC-LDPC code based 16QAM and 64QAM MLCM systems can achieve 0.4 dB and 1.2 dB net coding gain, respectively, compared with the recently proposed regular QC-LDPC code based 16QAM and 64QAM MLCM systems.

IRS辅助认知无线携能通信网络的发射功率最小化算法

智能反射平面(IRS)和认知无线携能通信技术被视为是提高能量效率和频谱利用率的潜在关键技术。文中研究了基于非线性能量采集模型的IRS辅助认知无线携能通信网络,其中次用户发射机同时给多个次用户接收机发送信息和能量,每个次用户接收机采用功率分割方式实现信息解码与能量采集,目的是通过联合优化次用户发射机的波束成形矢量、次用户接收机的功率分割系数以及IRS相移使次用户发射机的发射功率最小化。为了保证次用户发射机的信息与能量传输效率并限制次用户发射机对主用户接收机的同频干扰,考虑次用户接收机具有最小信干噪比约束、最小能量采集约束与功率分割系数约束,次用户发射机对主用户接收机有最大干扰功率值约束,以及IRS具有反射相移约束。所构建的问题属于非凸的二次约束二次规划问题,并且优化变量之间高度耦合,难以求解。文中提出一种基于半正定松弛法和连续秩一约束松弛法的交替优化算法进行高效求解。为了降低复杂度,进一步提出一种基于IRS分组的低复杂度优化算法。仿真结果表明,与几种基准算法相比,所提算法能够有效降低次用户发射机的发射功率。

融合时间信息的序列商品推荐模型

针对基于序列的推荐方法通常忽略用户的多种兴趣倾向,并且不能很好地获取用户在短期序列中的兴趣变化,从而导致推荐结果多样性不足的问题,提出了一种融合时间信息的序列商品推荐模型。首先,将用户的历史交互行为区分为短期序列与长期序列,分别采取不同的方法进行建模。对于短期序列,在传统的门控循环单元(GRU)结构中加入时间门,单独处理序列中的时间信息,同时利用多头自注意力机制捕获用户在同一会话中不同的兴趣方向;对于长期序列,采用DeepFM模型进行建模。最后,利用自适应的门控结构融合用户的长短期兴趣,并根据得到的兴趣向量计算商品的得分,排序后进行推荐。在淘宝数据集上的对比实验表明,该模型相较于主流的协同过滤模型,基于RNN、DNN的推荐模型以及BINN模型在命中率、平均倒数排名两个指标上都具有显著优势。

质子交换膜燃料电池健康状态指标融合

基于信息论的方法,筛选出包含较多燃料电池性能“不可逆衰退”信息的外部操作条件,采用二次规划法将筛选出的外部操作条件与输出电功率进行融合,提出了一种融合健康状态(state of health,SOH)指标。与以输出电功率为SOH的指标相比,采用融合SOH指标,在单调性、单池衰退一致性和综合评价指标上分别提升了22.83%、91.96%和55.60%。

基于SSA-MIC-SMBO-ESN的水质预测模型

水污染事件威胁人类生产和生活安全,提前预测水质变化对水污染的快速反应具有重要意义。基于水质数据的时序性,引入泄漏积分型回声状态网络(ESN),以莆田市东圳水库水质监测站的10种水质指标数据作为样本,分别构建DO、COD_(Mn)、TP的水质预测模型。首先,在用邻近点线性趋势法对缺失值进行填充,用Z-score法和邻近点线性趋势法对异常值进行检测修正的基础上,用奇异谱分析(SSA)算法对水质数据进行平滑降噪处理;然后,采用最大互信息系数(MIC)衡量水质指标之间的相关度,选取相关系数较大的水质指标作为待预测水质指标的输入特征;最后,利用ESN构建基于多特征的水质预测模型,其中采用序列模型优化(SMBO)算法对模型的超参数进行优化。试验结果表明,构建的DO、COD_(Mn)和TP的SSA-MIC-SMBO-ESN水质预测模型都具有较高的预测精度,适合实际应用。

多传感器非线性系统的序贯观测融合扩展卡尔曼滤波器

针对多传感器非线性系统,提出了一种序贯观测融合扩展卡尔曼滤波器(EKF)算法,其主要思想是根据传感器观测数据到达滤波器的先后次序依次进行融合处理。研究表明,提出的序贯观测融合EKF算法与传统的最优集中式观测融合EKF算法估计精度相当,且具有更低的计算复杂度,目标跟踪系统的仿真验证了所提算法的有效性。

基于自注意力网络的时间感知序列化推荐

随着信息技术的发展,推荐系统作为信息过载时代的重要工具,正扮演着越来越重要的角色.基于内容和协同过滤的传统推荐系统,倾向于以静态方式对用户与商品交互进行建模,以获取用户过去的长期偏好.考虑到用户的偏好往往是动态的,且具有非持续性和行为依赖性,序列化推荐方法将用户与商品的交互历史建模为有序序列,能有效捕获商品的依赖关系和用户的短期偏好.然而多数序列化推荐模型过于强调用户-商品交互的行为顺序,忽视了交互序列中的时间信息,即隐式假设了序列中相邻商品具有相同的时间间隔,在捕捉包含时间动态的用户偏好上具有局限性.针对以上问题,文中提出基于自注意力网络的时间感知序列化推荐(self-attention-based network for time-aware sequential recommendation,SNTSR)模型,该模型将时间信息融入改进的自注意力网络中,以探索动态时间对下一商品预测的影响.同时,SNTSR独立计算位置相关性,以消除可能引入的噪声相关性,增强捕获用户序列模式的能力.在两个真实世界数据集上的大量实验表明,SNTSR始终优于一组先进的序列化推荐模型.

分层多尺度决策信息系统的序贯三支决策

传统信息系统中每个属性只有一个尺度,但随着海量数据的涌现,实际应用中经常是在多个尺度上处理和分析问题.三支决策是解决分类问题的一种经典方法,序贯三支决策是在三支决策的基础上进行多步决策的一种方法.将多尺度决策信息系统与三支决策相结合,基于决策理论粗糙集提出分层多尺度决策信息系统的序贯三支决策模型,得到动态变化的正域、负域、边界域.对多尺度决策信息系统进行分层,依次在分层后得到的多个单尺度决策信息系统上进行讨论,构建尺度层面的序贯结构;在每个单尺度决策信息系统上,通过增加属性的方式得到属性子集序列,诱导出多级粒度结构,构建该尺度下粒度层面的序贯结构.为此,给出两种属性子集序列的选择方法;在序贯三支决策过程中,利用相对损失函数计算阈值,并讨论了阈值的性质;最后给出序贯三支决策过程中的分类规则,并用实例说明提出的模型能有效地处理分类问题.

地理距离是否影响金融机构决策——基于资产证券化市场微观数据的证据

通过建立贷款人与投资人之间的动态博弈模型,分析地理距离对贷款人、投资人决策的影响机制,并基于证券化产品底层基础资产数据进行验证。研究表明,贷款人倾向对与其地理距离较大的借款人开展更多的风险评估,同时倾向持有优质贷款、出售劣质资产和混合资产,导致投资人采取逆向选择策略,购买与贷款人地理距离较小的借款人的贷款。对于资质良好的贷款人,投资人会侧重评估基础资产风险;而对于资质较弱的贷款人,则会侧重贷款人的主体信用风险。为防范金融风险、促进证券化市场健康发展,应通过监管指标引导属地经营较强的中小银行开展证券化业务,建立更完善、更标准化的信息披露体系,提高大型银行的基础资产信息、中小银行和非银机构主体的信息披露程度。

Quantifying predictability of sequential recommendation via logical constraints

The sequential recommendation is a compelling technology for predicting users’next interaction via their historical behaviors.Prior studies have proposed various methods to optimize the recommendation accuracy on different datasets but have not yet explored the intrinsic predictability of sequential recommendation.To this end,we consider applying the popular predictability theory of human movement behavior to this recommendation context.Still,it would incur serious bias in the next moment measurement of the candidate set size,resulting in inaccurate predictability.Therefore,determining the size of the candidate set is the key to quantifying the predictability of sequential recommendations.Here,different from the traditional approach that utilizes topological constraints,we first propose a method to learn inter-item associations from historical behaviors to restrict the size via logical constraints.Then,we extend it by 10 excellent recommendation algorithms to learn deeper associations between user behavior.Our two methods show significant improvement over existing methods in scenarios that deal with few repeated behaviors and large sets of behaviors.Finally,a prediction rate between 64%and 80%has been obtained by testing on five classical datasets in three domains of the recommender system.This provides a guideline to optimize the recommendation algorithm for a given dataset.