Research on Bayesian Network Based User's Interest Model

It has very realistic significance for improving the quality of users＇ accessing information to filter and selectively retrieve the large number of information on the Internet. On the basis of analyzing the existing users＇ interest models and some basic questions of users＇ interest （representation, derivation and identification of users＇ interest）, a Bayesian network based users＇ interest model is given. In this model, the users＇ interest reduction algorithm based on Markov Blanket model is used to reduce the interest noise, and then users＇ interested and not interested documents are used to train the Bayesian network. Compared to the simple model, this model has the following advantages like small space requirements, simple reasoning method and high recognition rate. The experiment result shows this model can more appropriately reflect the user＇s interest, and has higher performance and good usability.

BP Network Based Users’Interest Model in Mining WWW Cache

By analyzing the WWW Cache model, we bring forward a user-interest description method based on the fuzzy theory and user-interest inferential relations based on BP（baek propagation） neural network. By this method, the users＇ interest in the WWW cache can be described and the neural network of users＇ interest can be constructed by positive spread of interest and the negative spread of errors. This neural network can infer the users＇ interest. This model is not the simple extension of the simple interest model, but the round improvement of the model and its related algorithm.

Term Structure of Interest Rates Based on Artificial Neural Network

In light of the nonlinear approaching capability of artificial neural networks （ ANN）, the term structure of interest rates is predicted using The generalized regression neural network （GRNN） and back propagation （BP） neural networks models. The prediction performance is measured with US interest rate data. Then, RBF and BP models are compared with Vasicek＇s model and Cox-Ingersoll-Ross （CIR） model. The comparison reveals that neural network models outperform Vasicek＇s model and CIR model, which are more precise and closer to the real market situation.

Geo-Social Profile Matching Algorithm for Dynamic Interests in Ad-Hoc Social Network

Among mobile users, ad-hoc social network (ASN) is becoming a popular platform to connect and share their interests anytime anywhere. Many researchers and computer scientists investigated ASN architecture, implementation, user experience, and different profile matching algorithms to provide better user experience in ad-hoc social network. We emphasize that strength of an ad-hoc social network depends on a good profile-matching algorithm that provides meaningful friend suggestions in proximity. Keeping browsing history is a good way to determine user’s interest, however, interests change with location. This paper presents a novel profile-matching algorithm for automatically building a user profile based on dynamic GPS (Global Positing System) location and browsing history of users. Building user profile based on GPS location of a user provides benefits to ASN users as this profile represents user’s dynamic interests that keep changing with location e.g. office, home, or some other location. Proposed profile-matching algorithm maintains multiple local profiles based on location of mobile device.

Visualization of Personal Interest Graph from Social Network

The advent of the time of big data along with social networks makes the visualization and analysis of networks information become increasingly important in many fields. Based on the information from social networks, the idea of information visualization and development of tools are presented. Popular social network micro-blog （＇Weibo＇） is chosen to realize the process of users＇ interest and communications data analysis. User interest visualization methods are discussed and chosen and programs are developed to collect users＇ interest and describe it by graph. The visualization results may be used to provide the commercial recommendation or social investigation application for decision makers.

Profile Matching in Electronic Social Networks Using a Matching Measure for Fuzzy Numerical Attributes and Fields of Interests

The problem of profile matching in electronic social networks asks to find those offering profiles of actors in the network fitting best to a given search profile. In this article this problem is mathematically formulated as an optimization problem. For this purpose the underlying search space and the objective function are defined precisely. In particular, data structures of search and offering profiles are proposed, as well as a function measuring the matching of the attributes of a search profile with the corresponding attributes of an offering profile. This objective function, given in Equation (29), is composed of the partial matching degrees for numerical attributes, discrete non-numerical attributes, and fields of interests, respectively. For the matching degree of numerical profile attributes a fuzzy value approach is presented, see Equation (22), whereas for the matching degree of fields of interest a new measure function is introduced in Equation (26). The resulting algorithm is illustrated by a concrete example. It not only is applicable to electronic social networks but also could be adapted for resource discovery in grid computation or in matchmaking energy demand and supply in electrical power systems and smart grids, especially to efficiently integrate renewable energy resources.

命名数据网络中基于包标记的Interest泛洪攻击缓解研究

命名数据网络因其关注请求对象本身而非地址并具有网间缓存等特点,得到了学术界的肯定.但在Interest泛洪攻击中,攻击者恶意占用PIT表等资源,导致其拒绝对合法用户服务,从而使网络遭受严重危害.针对基于熵的Interest泛洪攻击防御方案在定位攻击源、网络开销方面存在的不足,提出了一种基于包标记的缓解方法.该方法通过让Interest包携带边缘路由器信息,在检测到攻击并找出恶意前缀后对攻击源进行定位,然后向下游路由器发送溯源数据包,从而对攻击者采取限制措施.仿真结果表明:该方法可以更加精确地定位攻击源并有效地降低网络中的开销.

命名数据网络中基于信息熵的Interest洪泛攻击检测与防御

在命名数据网络中,兴趣包洪泛攻击通过向网络发送大量恶意interest包来消耗网络资源,从而对NDN造成较大危害.针对目前所提出的IFA攻击检测与防御方法存在攻击模式单一、在应对复杂攻击模式时效果不明显等局限,提出一种基于信息熵的改进方法(EIM),该方法通过与NDN路由器相连的用户的信誉值和信息熵相结合来限制攻击者发送的恶意interest包,很好地解决了现有方法在应对复杂的攻击模式时的局限性.仿真结果表明EIM较信息熵方法能够更有效地缓解IFA.

命名数据网络中基于增强隔离林的Interest包洪泛攻击检测方法

IFA是NDN中的恶意攻击之一,危害较大.针对IFA,提出基于增强隔离林的IFA检测方法,通过在构造增强隔离林的过程中区分Interest包所携带的合法前缀与异常前缀,对异常前缀进行进一步判断,从而准确地检测出恶意前缀.所提出的方法将检测出的恶意前缀列入黑名单中,限制携带恶意前缀的Interest包转发.仿真结果表明:该方法能够有效地缓解IFA.

命名数据网络中Interest洪泛攻击检测与防御

根据命名数据网络使用数据名称进行数据转发的特性,目前NDN中出现了一种基于兴趣包洪泛攻击的分布式拒绝服务攻击,对NDN网络危害较大.针对IFA,提出了一种基于Poseidon的改进方法——PAP,检测NDN路由器接口是否存在IFA,并构建interest包超时未响应表判断疑似恶意前缀.若存在IFA,则概率性限制该接口接收名称包含疑似恶意前缀的interest包速率,仿真结果表明PAP能够有效缓解IFA.

Enhancing Interest Forwarding for Fast Recovery from Unanticipated Data Access Failure in NDN

We show that an aggregated Interest in Named Data Networking (NDN) may fail to retrieve desired data since the Interest previously sent upstream for the same content is judged as a duplicate one and then dropped by an upstream node due to its multipath forwarding. Furthermore, we propose NDRUDAF, a NACK based mechanism that enhances the Interest forwarding and enables Detection and fast Recovery from such Unanticipated Data Access Failure. In the NDN enhanced with NDRUDAF, the router that aggregates the Interest detects such unanticipated data access failure based on a negative acknowledgement from the upstream node that judges the Interest as a duplicate one. Then the router retransmits the Interest as soon as possible on behalf of the requester whose Interest is aggregated to fast recover from the data access failure. We qualitatively and quantitatively analyze the performance of the NDN enhanced with our proposed NDRUDAF and compare it with that of the present NDN. Our experimental results validate that NDRUDAF improves the system performance in case of such unanticipated data access failure in terms of data access delay and network resource utilization efficiency at routers.

基于法益展开的非法利用信息网络罪“预备行为实行化”性质检视

面对日益严峻的网络违法犯罪活动,预防刑法观下的刑法立法呈现出积极扩张的一面。非法利用信息网络罪的增设所具备的“预备行为实行化”立法性质得到广泛认同,其通过对预备性质的网络行为独立处罚来实现刑法法益保护介入时点的提前。但这种网络犯罪立法中的“预备行为实行化”在规范与实践中又呈现出针对下游行为“适当扩张化”的新特性,由此直接造成犯罪圈层不当扩大及预备犯传统理论与司法解释不协调的问题。对此,应结合网络时代特质,明确非法利用信息网络罪是通过规制累积危险行为来直接保护网络空间秩序这一抽象化的独立集体法益,并在此基础上最终导向下游行为所面向的法益保护。应在法益二元论的立场下,以具有宪法基础的、注重社会安全利益积极保障的集体法益立法批判机能来证成本罪扩张性的一面,再以造成独立集体法益实质侵害这一解释规制机能的要求来对此进行限缩;还应坚持刑法最后法定位,对刑法辅助性法益保护局限性加以补充,为相应规范及理论进行注解。

Proposed Caching Scheme for Optimizing Trade-off between Freshness and Energy Consumption in Name Data Networking Based IoT

Over the last few years, the Internet of Things (IoT) has become an omnipresent term. The IoT expands the existing common concepts, anytime and anyplace to the connectivity for anything. The proliferation in IoT offers opportunities but may also bear risks. A hitherto neglected aspect is the possible increase in power consumption as smart devices in IoT applications are expected to be reachable by other devices at all times. This implies that the device is consuming electrical energy even when it is not in use for its primary function. Many researchers’ communities have started addressing storage ability like cache memory of smart devices using the concept called—Named Data Networking (NDN) to achieve better energy efficient communication model. In NDN, memory or buffer overflow is the common challenge especially when internal memory of node exceeds its limit and data with highest degree of freshness may not be accommodated and entire scenarios behaves like a traditional network. In such case, Data Caching is not performed by intermediate nodes to guarantee highest degree of freshness. On the periodical updates sent from data producers, it is exceedingly demanded that data consumers must get up to date information at cost of lease energy. Consequently, there is challenge in maintaining tradeoff between freshness energy consumption during Publisher-Subscriber interaction. In our work, we proposed the architecture to overcome cache strategy issue by Smart Caching Algorithm for improvement in memory management and data freshness. The smart caching strategy updates the data at precise interval by keeping garbage data into consideration. It is also observed from experiment that data redundancy can be easily obtained by ignoring/dropping data packets for the information which is not of interest by other participating nodes in network, ultimately leading to optimizing tradeoff between freshness and energy required.

诉求统筹下城中村改造利益相关者协同治理——以天津市北辰区改造项目为例

利益相关者协同治理是推动城中村改造项目顺利高效实施的关键,但改造实践中各阶段利益相关者关系及利益诉求错综复杂。以天津市北辰区改造项目为例,采用社会网络分析方法,系统、全面地分析项目各阶段利益相关者关系及其诉求的网络特征、演化规律。结果表明,拆迁安置补偿阶段利益相关者利益诉求协调难度最高;同一利益主体在项目不同阶段的角色功能差异明显;建设公众满意的生活环境是多元利益相关者的核心利益诉求。采用社会网络分析方法,剖析利益相关者关系及诉求协调的重点问题,为城中村改造利益相关者协同治理提供科学参考与依据。

一种融合表征的农产品推荐算法

针对农产品电商平台,产品季节性强、地域性强、用户行为多变,导致推荐效果不理想的问题,提出了一种融合表征的农产品推荐算法。首先,用长短期记忆网络和注意力网络相结合组成深度兴趣网络,以此来捕获物品的潜在特征;其次,构建用户-商品二部图;再次,利用图神经网络提取图数据的连接信息对每个节点的影响,并更新节点的嵌入式表示,以获取用户的潜在特征;最后,将两种潜在特征通过多层感知机得到待推荐农产品的购买概率,进一步提取和利用了用户行为序列中的用户深度兴趣,并将其融合深度兴趣网络进行推荐。实验结果表明:融合表征的农产品推荐算法相较于原有模型AUC指标提高9%以上,准确率和召回率提高约6%以上;相较于不考虑节点嵌入式表示的情况,AUC和准确率、召回率也均有提高。

融合自适应周期与兴趣量因子的轻量级GCN推荐

推荐系统在成熟的数据挖掘技术推动下,已能高效地利用评分数据、行为轨迹等显隐性信息,再与复杂而先进的深度学习技术相结合,取得了很好的效果.同时,其应用需求也驱动着对基础数据的深度挖掘与利用,以及对技术要求的减负成为一个研究热点.基于此,提出一种利用GCN(graph convolutional network)方法进行深度信息融合的轻量级推荐模型LG_APIF.该模型结合行为记忆,通过艾宾浩斯遗忘曲线模拟用户兴趣变化过程,采用线性回归等相对轻量的传统方法挖掘项目的自适应周期等深度信息;分析用户当前的兴趣分布,计算项目的兴趣量,以获取用户的潜在兴趣类型;构建用户-类型-项目三元组的图结构,并结合减负后的GCN技术来生成最终的项目推荐列表.实验验证所提方法的有效性,通过与8个经典模型在Last.fm,Douban,Yelp,MovieLens数据集中的对比,表明该方法在Precision,Recall及NDCG指标上都得到良好改善,其中,Precision平均提升2.11%,Recall平均提升1.01%,NDCG平均提升1.48%.

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%。

古典小说叙事志趣的复兴:网络文学与IP电影的“情”“志”表达

近年来,消费文化盛行,以短视频、博客段子等为代表的碎片化、娱乐化的视觉狂欢文化逐渐代替传统的文字、图像艺术成为中国社会典型的后现代文化症候。网络文学生成于这样的大环境之中,其作品改编IP电影也随着行业的鼎沸成为一股流行趋势。然而市场喧嚣之余,大量创作者却始终保有一定的“寄怀”思绪,通过对外“感物”和向内自醒的咏志抒情表达,有意或无意地成为传统叙事志趣与现代人文意识之间的桥梁,在传承与复苏中国叙事志趣传统的同时,为当代通俗文化与大众美学赋予更细腻的情感结构与更多的文化内涵。

社交参与视角下超图增强的学习趣缘社群群体检测研究

在线学习群体检测是在新一轮科技革命赋能教育创新变革背景下,依据学习者个性化特征优化教育资源分层配置的关键途径。现有学习趣缘社群在线学习群体的检测主要依赖学习者的直接行为记录和互动指标,较少关注学习者潜在的社交参与水平和社群结构。为营造数智环境下学习者画像决策辅助全民自主学习的文化氛围,本文提出一种社交参与视角下超图增强的学习趣缘社群群体检测方法。首先,从影响用户社交参与的维度出发,构建能够体现学习者社交参与水平的特征集。其次,提出超图卷积网络(hypergraph convolutional network,HyperGCN)增强的图聚类算法HG-SDCN(structural deep clustering network based on HyperGCN),解决了利用二分图检测在线学习群体时无法有效捕捉学习者多元交互关系和高阶结构的问题。最后,从真实学习趣缘社群收集数据,验证本文提出方法的检测效果。与基线相比,本文方法在Acc(accuracy)、F1、NMI(normalized mutual information)和ARI(adjusted Rand index)等评价指标上分别提升了16.16、9.77、16.01和22.14个百分点。上述结果不仅证明了HyperGCN在捕捉学习者高阶结构实现在线学习群体检测任务中的有效性,还为未来从社交参与维度制定调整个性化教育资源配置策略提供了方法和理论支撑。

融合时间感知和多兴趣提取网络的序列推荐

针对序列推荐任务中的时间动态性和多重兴趣建模问题,提出一种时间感知的项目嵌入方法,用于学习项目之间的时间关联性。在此基础上,提出一种融合时间感知和多兴趣提取网络的序列推荐(time-aware multi-interest sequence recommendation,TMISA)方法。TMISA采用自注意力序列推荐模型作为局部特征学习模块,以捕捉用户行为序列中的动态偏好;通过多兴趣提取网络对用户的全局偏好进行建模;引入门控聚合模块将局部和全局特征表示动态融合,生成最终的用户偏好表示。实验证明,在5个真实推荐数据集上,TMISA模型表现出卓越性能,超越了多个先进的基线模型。