基于位置聚类和张量分解的Web服务推荐算法

基于服务质量（QoS）的Web服务推荐能在众多功能相似的Web服务中发现最能满足用户非功能需求的Web服务,但QoS属性值预测算法仍存在预测准确度不高和数据稀疏性的问题。针对以上问题,提出了一种基于位置聚类和分层张量分解的QoS预测算法ClustTD,该算法基于用户和服务的位置属性将用户和服务聚类成多个局部组,分别对局部组和全局的用户、服务和时间上下文进行张量建模和分解,将局部和全局张量分解的QoS预测值进行加权组合,同时考虑了局部和全局因素,获得最终QoS预测值。实验结果表明,该算法具有较高的QoS预测准确率和Web服务推荐质量,并能在一定程度上解决数据稀疏性问题。     

A novel methodology for prediction of spatial-temporal activities using latent features

In today's era of big data, huge amounts of spatial-temporal data are generated daily from all kinds of citywide infrastructures. Understanding and predicting accurately such a large amount of data could benefit many real-world applications. In this paper, we propose a novel methodology for prediction of spatial-temporal activities such as human mobility, especially the inflow and outflow of people in urban environments based on existing large-scale mobility datasets. Our methodology first identifies and quantifies the latent characteristics of different spatial environments and temporal factors through tensor factorization. Our hypothesis is that the patterns of spatial-temporal activities are highly dependent on or caused by these latent spatial-temporal features. We model this hidden dependent relationship as a Gaussian process, which can be viewed as a distribution over the possible functions to predict human mobility. We tested our proposed methodology through experiments conducted on a case study of New York City's taxi trips and focused on the mobility patterns of spatial-temporal inflow and outflow across different spatial areas and temporal time periods. The results of the experiments verify our hypothesis and show that our prediction methodology achieves a much higher accuracy than other existing methodologies.     

基于模体演化的时序链路预测方法

时序链路预测是动态网络分析的重要组成部分,具有极大的理论和应用价值. 传统的时序链路预测方法往往直接对边的演化规律进行分析,忽略了网络中其他微观结构的演化对链路形成的影响. 基于此分析,本文引入非负张量分解和时间序列分析对网络模体的演化规律进行研究,进而提出一种基于模体演化的链路预测方法. 在三个真实数据集上的实验结果表明,该方法能有效提高链路预测精度.     

Where to go?Predicting next location in IoT environment

Next location prediction has aroused great inter-ests in the era of internet of things(IoT).With the ubiquitous deployment of sensor devices,e.g..GPS and Wi-Fi,loT en-vironment offers new opportunities for proactively analyzing human mobility patterns and predicting user's future visit in low cost,no matter outdoor and indoor.In this paper,we con-sider the problem of next location prediction in loT environ-ment via a session-based manner.We suggest that user's future intention in each session can be better inferred for more ac-curate prediction if patterns hidden inside both trajectory and signal strength sequences ollected from IoT devices can be jointly modeled,which however existing state-of the-art meth-ods have rarely addressed.To this end,we propose a trajectory and sIgnal sequence(TSIS)model,where the trajectory transi-tion regularities and signal temporal dynamics are jointly embedded in a neural network based model.Specifically,we employ gated recurrent unit(GRU)for capturing the temporal dy-namics in the mutivariate signal strength sequence.Moreover,we adapt gated graph neural networks(gated GNNs)on loca-tion transition graphs to explicitly model the transition patterns of trajectories.Finally,both the low-dimensional representa-tions learned from trajectory and signal sequence are jointly optimized to construct a session embedding,which is further employed to predict the next location.Extensive experiments on two real-world Wi-Fi based mobility datasets demonstrate that TSIS is effective and robust for next location prediction pompared with other competitive baselines.     

A PID-incorporated Latent Factorization of Tensors Approach to Dynamically Weighted Directed Network Analysis

A large-scale dynamically weighted directed network(DWDN)involving numerous entities and massive dynamic interaction is an essential data source in many big-data-related applications,like in a terminal interaction pattern analysis system(TIPAS).It can be represented by a high-dimensional and incomplete(HDI)tensor whose entries are mostly unknown.Yet such an HDI tensor contains a wealth knowledge regarding various desired patterns like potential links in a DWDN.A latent factorization-of-tensors(LFT)model proves to be highly efficient in extracting such knowledge from an HDI tensor,which is commonly achieved via a stochastic gradient descent(SGD)solver.However,an SGD-based LFT model suffers from slow convergence that impairs its efficiency on large-scale DWDNs.To address this issue,this work proposes a proportional-integralderivative(PID)-incorporated LFT model.It constructs an adjusted instance error based on the PID control principle,and then substitutes it into an SGD solver to improve the convergence rate.Empirical studies on two DWDNs generated by a real TIPAS show that compared with state-of-the-art models,the proposed model achieves significant efficiency gain as well as highly competitive prediction accuracy when handling the task of missing link prediction for a given DWDN.     

Analyzing movement predictability using human attributes and behavioral patterns

The ability to predict human mobility, i.e., transitions between a user's significant locations (the home, workplace, etc.) can be helpful in a wide range of applications, including targeted advertising, personalized mobile services, and transportation planning. Most studies on human mobility prediction have focused on the algorithmic perspective rather than on investigating human predictability. Human predictability has great significance, because it enables the creation of more robust mobility prediction models and the assignment of more accurate confidence scores to location predictions. In this study, we propose a novel method for detecting a user's stay points from millions of GPS samples. Then, after detecting these stay points, a long short-term memory (LSTM) neural network is used to predict future stay points. We explore the use of two types of stay point prediction models (a general model that is trained in advance and a personal model that is trained over time) and analyze the number of previous locations needed for accurate prediction. Our evaluation on two real-world datasets shows that by using our preprocessing approach, we can detect stay points from routine trajectories with higher accuracy than the methods commonly used in this domain, and that by utilizing various LSTM architectures instead of the traditional Markov models and advanced deep learning models, our method can predict human movement with high accuracy of more than 40% when using the Acc@1 measure and more than 59% when using the Acc@3 measure. We also demonstrate that the movement prediction accuracy varies for different user populations based on their trajectory characteristics and demographic attributes.     

基于贝叶斯张量分解的改进QoS预测

随着Web服务相关标准和技术的日趋成熟,基于服务质量（QoS）的Web服务推荐对用户体验起着决定性作用。如何准确预测Qos值是当今的研究热点。以往基于近邻或模型的协同过滤算法,采用的是“用户-服务”二维信息,预测的QoS值是静态的且精准性不高。将时间信息维度引入张量模型,建立“用户-服务-时间”的三维张量可使QoS预测值更加符合用户需求特点,用贝叶斯方法求解张量分解,引入概率意义下对于系统的解释和分析,提供一套先验概率引入先验知识的贝叶斯推断框架,提高了QoS预测的精确度。实验表明,使用该算法的预测结果较其他算法相比较有更小的平均绝对误差,很好地解决了数据稀疏度问题。     

基于张量分析的链接聚类算法的研究*

针对复杂信息网络中多链接高维数据聚类难以处理且效率较低问题,提出了一种新颖的基于高阶张量分析方法和模块化网络分析方法相结合的链接聚类算法。利用模块化方法分析网络,利用张量的形式表示多维的复杂的多链接数据,利用Tucker张量分解的方法对数据降维处理,降低了算法的时间和空间复杂度。并在复杂网络环境下,通过实验验证了算法的有效性和健壮性。     

基于模式匹配度的用户移动规则挖掘及位置预测方法研究

传统的移动用户位置预测方法由于模式支持度计算方式不合理,存在预测精度偏低的问题。为此,提出了一种基于模式匹配度的用户移动规则挖掘及位置预测方法,并将其用于移动通信系统中,以基台覆盖范围网格为单元的用户位置预测。具体包括三个步骤：通过图的遍历挖掘用户移动模式、基于用户移动模式生成用户移动规则和依据用户移动规则进行位置预测。实验分析使用10个批次轨迹数据进行用户移动规则挖掘,结果表明,该方法挖掘出的用户移动规则数少、支持度高和置信度高,具有高精度的优点。     

轴承在线早期故障检测的无监督张量深度迁移学习方法

基于张量分解可有效挖掘信号高维本质信息的优点,提出一种无监督张量深度迁移学习方法.首先,构建基于张量表示的深度多任务异常检测模型,利用核心张量构建单分类异常检测规则表示,并建立超球规则适配机制,交替优化张量分解和域无关特征提取,以实现异常检测规则在离线轴承和在线目标轴承间的有效传递,完成在线无标记数据的异常检测;其次,提出一个基于异常概率贯序累积的非参数报警阈值设定方法,可在仅设定误报警率置信度的条件下自适应选择在线阈值,并给出该阈值合理性的理论分析.在IEEE PHM Challenge 2012轴承数据集上进行实验,结果表明,所提出方法可获得更好的检测实时性和更低的误报警数,为早期故障检测提供一种具有易部署性和鲁棒性的解决方案.     

核张量子空间分解EEG特征提取方法研究

针对共空间模式（Common Spatial Patterns,CSP）对源信号和记录的脑电信号之间严格的线性模式的假设关系,充分发挥张量在多维上同时处理的优势,研究了一种核张量子空间分解EEG特征提取方法。首先生成EEG数据的张量,利用带二次等式约束的最小二乘问题解决张量分解问题,并将张量扩展到子空间,减小计算的压力,最后推广到核空间,将数据投影到高维特征空间来增强辨别能力。实验数据采用2005年BCI竞赛III的数据集III_3a,实验结果表明,KTSD方法能够从多类运动想象任务的EEG数据中提取相应的特征,并得到较好分类结果和运行效率。     

基于超图正则化非负Tucker分解的图像聚类算法

针对非负张量分解应用于图像聚类时忽略了高维数据内部几何结构的问题,在经典的张量非负Tucker分解的基础上,添加超图正则项以尽可能多地保留原始数据的内在几何结构信息,提出一种基于超图正则化非负Tucker分解模型HGNTD。通过构造超图刻画数据内部样本间的高阶关系,提高几何结构描述的准确性,针对超图正则化非负张量分解模型,基于交替非负最小二乘法,设计快速有效的超图正则化非负Tucker分解算法求解所给模型,证明算法在非负的条件下是收敛的,最终将算法应用于图像聚类。在Yale和COIL两个常用公开数据集上的实验结果表明,相对于k-means、非负矩阵分解、图正则化非负矩阵分解、非负Tucker分解和图正则化非负Tucker分解等算法,超图正则化非负Tucker分解算法聚类准确度提升了8.6%~11.4%,归一化互信息提升了2.0%~7.5%,具有更好的聚类效果。     

融合时空特征的光伏气象因子预测模型

光伏发电极易受到天气的影响而具有波动性和不确定性,因此对气象因子的准确预测对光伏电站的运维具有重要意义。提出了一种基于深度学习的时空特征融合模型,实现对光伏气象因子的精准预测。在时间维度上,设计了一种改进的长短期记忆模块,融合注意力机制和遗传算法,得到最优注意力参数以提高预测精度;在空间维度上,将光伏电站所在区域按照经纬度划分,利用张量分解对区域内气象因子进行预测。在中国东南部某光伏系统的真实数据集上,对该模型的有效性进行了评估。结果表明,该模型在时间维度和空间维度均具有较高预测精度,同时对稀疏数据有较强的鲁棒性。     

基于EMD-GAELM-ARIMA算法的大坝变形预测

In view of the fact that it is difficult for statistical models to make good predictions of nonlinear and non-stationary dam deformation, artificial intelligence algorithms are induced. The empirical mode decomposition method (EMD), genetic algorithm (GA) optimized extreme learning machine (ELM), and ARIMA error correction model were used to construct a dam deformation prediction model. First this paper uses EMD to decompose and reconstruct the monitoring data to stabilize it and obtain eigenmode functions and residual sequences with physical significance; then uses GAELM to analyze and predict the decomposition results; finally, uses ARIMA model to correct errors. Taking a concrete rockfill dam as an example, the dam deformation prediction model constructed by the optimization algorithm is used to analyze and predict it. The analysis results show that the EMD-GAELM-ARIMA model algorithm has higher prediction accuracy than the traditional single algorithm. It is feasible in dam deformation prediction.     

基于气象数据可视化降维和多模加权组合的短期负荷预测

针对短期负荷预测中数据预处理的必要性和单一预测模型的局限性,提出了一种基于气象数据可视化降维和多模加权组合的短期负荷预测方法。该方法将可视化降维、模态分解降噪、单一预测模型和权重确定理论相结合,构建了气象数据降维、历史负荷分解、模态分量降噪和多模加权组合的短期负荷预测模型。通过设置3种对比实验环境,对某地区供电公司所提供的电力负荷和气象数据进行分析。预测结果及误差分析表明,所提短期负荷预测方法在保留高维气象因素本质特征结构的同时,能有效结合数据预处理方法及单一预测模型的特点,有效提升该地区电网负荷的预测精度。     

基于LightG BM 的高校就业预测模型

In view of the characteristics of high-dimensional, unbalanced and multi-category employment data, in order to further im- prove the accuracy of decision tree method in the employment prediction of college students, an employment prediction model based on LightGBM is proposed. First the improved ADASYN sampling algorithm is used to increase the minority class in the data sam- ple, and then the employment data after balance is used for training LightGBM algorithm, and Bayesian model is used for parameter optimization to get the final employment prediction. Finally the prediction model is analyzed to measure the influence of each fea- ture on employment. The validity of the proposed method is verified through the data set of unbalanced employment data of college graduates, and compared with various unbalanced classification methods. It is proved that the proposed model has better prediction performance.     

结合稀疏先验与多模式分解的低秩张量恢复方法

目的 各类终端设备获取的大量数据往往由于信息丢失而导致数据不完整,或经常受到降质问题的困扰。为有效恢复缺损或降质数据,低秩张量补全备受关注。张量分解可有效挖掘张量数据的内在特征,但传统分解方法诱导的张量秩函数无法探索张量不同模式之间的相关性;另外,传统张量补全方法通常将全变分约束施加于整体张量数据,无法充分利用张量低维子空间的平滑先验。为解决以上两个问题,提出了基于稀疏先验与多模式张量分解的低秩张量恢复方法。方法 在张量秩最小化模型基础上,融入多模式张量分解技术以及分解因子局部稀疏性。首先对原始张量施加核范数约束,以此捕获张量的全局低秩性,然后,利用多模式张量分解将整体张量沿着每个模式分解为一组低维张量和一组因子矩阵,以探索不同模式之间的相关性,对因子矩阵施加因子梯度稀疏正则化约束,探索张量子空间的局部稀疏性,进一步提高张量恢复性能。结果 在高光谱图像、多光谱图像、YUV（也称为YCbCr）视频和医学影像数据上,将本文方法与其他8种修复方法在3种丢失率下进行定量及定性比较。在恢复4种类型张量数据方面,本文方法与深度学习GP-WLRR方法（global prior refined weighted low-rank representation）的修复效果基本持平,本文方法的MPSNR（mean peak signal-to-noise ratio）在所有丢失率及张量数据上的总体平均高0.68dB,MSSIM（mean structural similarity）总体平均高0.01;与其他6种张量建模方法相比,本文方法的MPSNR及MSSIM均取得最优结果。结论 提出的基于稀疏先验与多模式张量分解的低秩张量恢复方法,可同时利用张量的全局低秩性与局部稀疏性,能够对受损的多维视觉数据进行有效修复。     

以用户为中心的最佳位置区域预测

阐述了具有蜂窝网络结构的个人通信系统中,基于考虑静止概率全向运动模型的动态位置管理。首先介绍了全向运动模型及其改进,使之更加符合实际用户的运动特性;在此基础上,提出了一种新型的以用户为中心的位置区域预测方法,通过计算各小区访问概率并选择其中的极大值,预测出以用户为中心的最佳位置区域;文中还提出仿真算法,并举例说明仿真结果。     

基于改进带偏置概率矩阵分解算法的研究

针对个性化推荐过程中高维稀疏性问题,本文提出一种将奇异值分解技术和带偏置概率矩阵分解相结合的推荐方法。 首先利用SVD算法初始化用户项目潜在因子向量,避免因随机赋值而使得函数陷入局部最优解,接着将用户项目的偏置信息融入到概率矩阵分解算法中,同时为了提升训练速度和推荐精度,通过动量加速的迷你批量梯度下降（mini Batch Gradient Descent,miniBGD）来训练,最后利用分解后的两个低维矩阵对原矩阵中的未知评分进行预测,在三个公开数据集的实验结果表明,本文提出的算法相对于传统的算法能够有效的提高推荐精度,进一步缓解由数据高维稀疏性带来的推荐质量不高的问题。     

基于自监督学习的社交网络用户轨迹预测模型

针对当前用户轨迹数据建模中存在的签到点稀疏性、长时间依赖性和移动模式复杂等问题,提出基于自监督学习的社交网络用户轨迹预测模型SeNext,对用户轨迹进行建模和训练来预测用户的下一个兴趣点（POI）。首先,使用数据增强的方式来丰富训练数据样本,以解决数据不足及个别用户足迹太少导致的模型泛化能力不足的问题;其次,将循环神经网络（RNN）、卷积神经网络（CNN）和注意力机制分别用于当前轨迹和历史轨迹的建模中,以此从高维稀疏的数据中提取有用的表示,用来匹配用户过去最相似的移动方式。最后,通过结合自监督学习并引入对比损失优化噪声对比估计（InfoNCE）,SeNext在潜在空间学习隐含表示来预测用户的下一个POI。实验结果表明,在纽约数据集上,SeNext比最新的VANext（Variational Attention based Next）模型的预测准确度在Top@1上提高了11.10%左右。