一种基于信任的协同过滤推荐模型

传统的协同过滤推荐技术主要基于用户-项目评价数据集进行挖掘推荐,没有有效地利用用户通信上下文信息,从而制约其进一步提高推荐的精确性。针对传统协同过滤推荐算法存在的推荐精度不高的弊端,在协同过滤算法中融入通信上下文信息,引入了通信信任、相似信任和传递信任三个信任度,并提出了一种基于信任的协同过滤推荐模型。通过公开数据集验证测试,证明提出的推荐算法较传统的协同过滤推荐技术在推荐准确性上有较大提高。     

基于模糊处理和上下文敏感SVM的协同过滤推荐新方法

协同过滤系统是电子商务最重要的技术之一,用户相似度算法的优劣直接决定推荐性能的准确程度。现有推荐方法忽略用户上下文特征,因而用户相似度判定较差。针对该问题,该文提出了FSVMCF方法,该方法采取模糊上下文数据及上下文敏感SVM和协同过滤相结合方法,提高了推荐准确度。该文实验以汽车、飞机、火车等交通工具图像作为推荐对象,实验结果验证了该方法对于图像推荐的性能有较大提高。     

基于多上下文信息的协同过滤推荐算法

随着电子商务和互联网的发展,数据信息呈爆炸式增长,协同过滤算法作为一种简单而高效的推荐算法,能在一定程度上有效地解决信息爆炸问题。但是传统协同过滤算法仅通过单一评分来挖掘相似用户,推荐效果并不占优势。为了提高个性化推荐的质量,如何充分利用用户(物品)的文本、图片、标签等上下文信息以使数据价值最大化是当前推荐系统亟待解决的问题。对此,提出了一种融合多种类型上下文信息的协同过滤算法。以用户商品交互信息为二部图,根据不同类型上下文的特点构建不同的相似度网络,设计目标函数在多种上下文信息网络的约束下联合矩阵分解,并学得用户商品的表示学习。在多个数据集上进行了充分实验,结果表明,融合多种类型上下文信息的协同过滤算法不仅能有效提高推荐的准确度,而且能在一定程度上解决数据稀疏性问题。     

社交网络中上下文感知协同过滤算法

围绕上下文感知推荐技术和社会化网络推荐技术的局限性展开研究,提出一种基于社会化网络环境下的名为HCCF的上下文感知协同过滤方法。在充分考虑上下文感知推荐系统实际问题的基础上,首先量化了不同维度的上下文对推荐系统所产生的影响,并在此基础上定义了上下文影响系数。在此基础上引入了社会化网络环境中不同用户之间的相互影响,并采用社会化网络用户信任度进行衡量,最后对上下文因素和社会化网络用户信任度进行综合考虑,提出一种新的相似度计算方法。理论分析和在真实数据集上的实验结果表明,相对于单纯基于上下文的系统过滤算法以及社会化网络推荐方法而言,该算法的准确性和推荐效率均得到一定程度的提升。     

基于社交用户标签的混合top-N推荐方法

针对协同过滤方法的冷启动问题,提出一种将社交用户标签与协同过滤相结合的混合top-N推荐方法。通过社交用户关系获得可信用户集,然后根据个性化标签采用结构上下文相似性算法(SimRank)计算社交用户相似近邻集并进行预测推荐,最后结合传统协同过滤方法进行推荐。实验结果表明,该方法能够提高在一般数据集及冷启动用户数据集下的推荐精度。     

Spark平台下基于上下文信息的影片混合推荐

响应速度较慢和推荐内容与用户上下文信息匹配程度低是当前影片推荐系统迫切需要解决的问题。针对上述挑战,提出Spark平台下基于上下文信息的影片混合推荐方法。它利用分布式并行计算技术Spark进行加速,来提高系统对于海量数据的检索与计算速度,从而减少了系统响应时间。同时该方法将“上下文推荐”和“交替最小二乘的协同过滤（ALS）”融合成一种混合推荐方法,提高了系统的推荐精度。实验结果表明,所提出的混合推荐方法有不错的效果。     

基于Hadoop的协同过滤推荐算法

在电子商务推荐系统中,协同过滤算法技术是主要的采用技术,而推荐系统的准确率受相似度方法的直接影响。本文通过对用户共同评分项的流行度和用户的特征之间的相关性的分析,给出了改进的相似度度量方法及相应的协同过滤推荐算法,设计并实现了基于Hadoop的协同过滤推荐系统。利用MAE分别对原始协同过滤推荐效果和改进后的推荐算法进行测评,结果表明,该算法在一定程度上提高了推荐的准确度。同时,对分布式推荐平台的整体性能实验表明,随着虚拟机节点的适度增加,协同过滤推荐引擎的计算时间不断减少,这表明推荐引擎的总体性能较传
　　统单机推荐引擎得到提升。     

基于遗传算法的聚类与协同过滤组合推荐算法

使用协同过滤进行推荐,在处理大数据集时存在效率问题和推荐结果质量不高的问题。k均值聚类在处理大数据集时有着较好的性能。针对使用协同过滤进行推荐存在的问题,通过使用遗传算法将聚类和协同过滤组合起来进行项目推荐,以此来提高推荐算法的推荐效率和推荐质量,降低组合聚类和协同过滤进行推荐的复杂度。使用组合得到的算法在MovieLens数据集上做推荐对比实验,结果表明,相比单纯使用协同过滤进行推荐,使用基于遗传算法的聚类与协同过滤组合推荐算法进行项目推荐,能得到质量更好的推荐结果。     

基于遗传算法的聚类与协同过滤组合推荐算法

使用协同过滤进行推荐,在处理大数据集时存在效率问题和推荐结果质量不高的问题。k均值聚类在处理大数据集时有着较好的性能。针对使用协同过滤进行推荐存在的问题,通过使用遗传算法将聚类和协同过滤组合起来进行项目推荐,以此来提高推荐算法的推荐效率和推荐质量,降低组合聚类和协同过滤进行推荐的复杂度。使用组合得到的算法在MovieLens数据集上做推荐对比实验,结果表明,相比单纯使用协同过滤进行推荐,使用基于遗传算法的聚类与协同过滤组合推荐算法进行项目推荐,能得到质量更好的推荐结果。     

基于Hadoop平台协同过滤推荐算法

针对协同过滤推荐算法在数据稀疏性及在大数据规模下系统可扩展性的两个问题, 在分析研究Hadoop分布式平台与协同过滤推荐算法后, 提出了一种基于Hadoop平台实现协同过滤推荐算法的优化方案. 实验证明, 在Hadoop平台上通过MapReduce结合Hbase数据库实现算法, 能够有效地提高协同过滤推荐算法在大数据规模下的执行效率, 从而能够进一步地搭建低成本高性能、动态扩展的分布式推荐引擎.     

基于协同虚拟环境的装配设计系统

针对分布式环境中复杂产品协同开发的装配设计需求,研究基于协同虚拟环境的装配设计系统的体系结构和关键技术,提出分布式条件下支持多用户实时交互操作的协同虚拟装配方法,包括复杂场景高分辨率实时渲染和多通道沉浸输出方案。以此为基础开发分布式协同虚拟装配环境,并通过面向汽车整车的多用户协同虚拟装配实例验证了系统的有效性。结果证明,用户可在该系统中完成复杂产品的协同虚拟装配设计工作。     

Hybrid client-server architecture and control techniques for collaborative product development using haptic interfaces

In this paper, a collaborative product development and prototyping framework is proposed by using distributed haptic interfaces along with deformable objects modeling. Collaborative Virtual Environment (CVE) is a promising technique for industrial product development and virtual prototyping. Network control problems such as network traffic and network delay in communication have greatly limited collaborative virtual environment applications. The problems become more difficult when high-update-rate haptic interfaces and computation intensive deformable objects modeling are integrated into CVEs for intuitive manipulation and enhanced realism. A hybrid network architecture is proposed to balance the computational burden of haptic rendering and deformable object simulation. Adaptive artificial time compensation is used to reduce the time discrepancy between the server and the client. Interpolation and extrapolation approaches are used to synchronize graphic and haptic data transmitted over the network. The proposed techniques can be used for collaborative product development, virtual assembly, remote product simulation and other collaborative virtual environments where both haptic interfaces and deformable object models are involved.     

地图协同标绘感知系统的设计与实现

传统标绘系统都是单用户标绘，它不能满足多用户同时进行标绘的要求。文中设计构建了一个CSCW标绘系统，以解决远程多用户同时进行协同标绘行为的难题，并提出主席一用户模式体系结构，以协同标绘层和协同管理层为核心的系统框架以及分布式加锁策略等技术来解决协同管理、实时交互和协同感知，以此实现异地协同标绘系统。     

A CAE-integrated distributed collaborative design system for finite element analysis of complex product based on SOOA

Large-scale finite element analysis of complex product needs a wider support from external CAE resources. This paper proposes a method of the distributed concurrent and collaborative design in the distributed intelligent resources environment to use Web-based extended manufacturing resources for product development. The CAE-integrated distributed collaborative design system architecture and enabling techniques are presented. Distributed CAE service resources on the Internet are achieved through Jini^TM and service object-oriented architecture technologies. These CAE software tool resources are encapsulated as service providers in the system. A case study of railway bogie development is presented, and the Pro/E, HyperMesh, Ansys and HumanExpert service providers dynamically invoked and integrated as a temporary service federated environment with no need to know the exact location of a provider beforehand, flexibility can be well achieved, and collaborative design in the system can be implemented in the distributed dynamic environment. Compared with the traditional design system, some preliminary results indicate this architecture can shorten the cycle of service exchange, and strongly support concurrent and collaborative design of the complex product.     

复杂产品虚拟样机支撑平台的初步研究与开发

该文主要研究并介绍了复杂产品虚拟样机支撑平台研究项目取得的阶段研究成果，包括支撑平台体系结构及其关键技术等，尤其重点介绍了目前自行开发集成的，开放的，基于标准支持多领域复杂产品虚拟样机工程的支撑平台，并给了两个基于支撑平台的典型虚拟样机开发应用案例。初步实践表明复杂产品虚拟样机支撑平台可看作为下一代企业协同应用各领域CAx/DFx工具的有效手段。最后，对进一步研究工作进行了展望。     

Multi-User Collaborative Assembly Planning over the Internet

1 Introduction Assembly planning is an important issue of product design for assembly, and many achievements have been obtained. Techniques such as geometric reasoning [1,2], genetic algorithms (GA) [3], virtual assembly [4,5] and so on are employed to assembly planning of relatively simple products successfully. But for the more complex cases, there is a combinational complexity problem when the geometric reasoning method is used. The fitness function in GA cannot take into account various…     

复杂产品虚拟样机技术研究

虚拟样机技术尤其是面向多领域支持复杂系统的协同虚拟样机仿真支撑平台的研究已成为研发的热点。通过对Agent技术和WEB服务技术以及虚拟样机仿真系统的长期研究,提出了一种面向多领域的复杂产品虚拟样机仿真支撑平台的实现方案,给出了该仿真支撑平台的基于Agent和WEB服务的系统结构和虚拟样机设计中涉及的机械,电子,控制,软件等领域的智能Agent的统一描述,讨论了虚拟样机仿真支撑系统中Agent间的通信方式并阐明了系统的工作流程和信息集成模式,为虚拟样机的进一步发展和应用奠定了基础。     

Multi-User Collaborative Assembly Planning over the Internet

Due to the increasing complexity of products and for the distributed product development, more closely collaborative work among designers is required. A collaborative assembly planning approach is proposed to support assembly planning in a networked environment. The working procedure is depicted and the key techniques including collaborative-planning-oriented assembly decomposition modeling, assembly assignment modeling, and sub-plans merging are addressed. By incorporating visual models at client side with assembly application models at server side, a web-based supporting environment for collaborative assembly planning has been developed using VRML and Java-EAI techniques. A case study is given to illustrate the feasibility and validity of the idea.     

A self-configurable large-scale virtual manufacturing environment for collaborative designers

As manufacturing environments are getting distributed and increasing in size, the related virtual environments are getting larger and more closely networked together. This trend has led to a new paradigm—large-scale virtual manufacturing environment (LSVME). It supports networked and distributed virtual manufacturing to meet manufacturing system requirements. Since it contains a large number of virtual components, an effective data structure and collaborative construction methodology are needed. A metaearth architecture is proposed as the data structure for representing LSVME. This architecture consists of virtual space layer, mapping layer, library layer and ontology layers, which describe interaction among virtual components and has the ability to analyze the characteristics of virtual environment. In addition, it increases reusability of virtual components and supports self-reconfiguration for manufacturing simulation. A heuristic construction method based on graph theory is proposed using this architecture. It prevents redundant design of virtual components and contributes to an effective construction scheduling technique for collaborative designers.     

A framework for collaborative top-down assembly design

This paper presents a new system framework for collaborative top-down assembly design. Different from current computer-aided design (CAD) systems, the framework allows a group of designers to collaboratively conduct product design in a top-down manner. In our framework, a multi-level and distributed assembly model is adopted to effectively support collaborative top-down assembly design. Meanwhile, fine-granularity collaborative design functionalities are provided. First, the coupled structural parameters involved in the distributed skeleton models of the product can be collaboratively determined by the correlative designers based on fuzzy and utility theory. Second, agent based design variation propagation is achieved to ensure the consistency of the multi-level and distributed assembly model during the whole design process. Third, collaborative design of assembly interfaces between the components assigned to different designers is supported. The prototype implementation shows that our framework works well for supporting practical collaborative top-down assembly design.