基于云计算的视频推荐系统的设计

通过对现有视频网站的调查研究,发现大部分都存在信息过载的问题。所以对视频网站来说拥有推荐系统是有必要的。通过对现有视频推荐系统的分析研究,利用开源云计算技术—Hadoop,及其部分相关组件Hive、Hbase等,设计了一种基于云计算的个性化视频推荐系统,此系统仅适用于以专业视频为主的网站。     

基于位置服务的分布式差分隐私推荐方法研究

随着移动互联网技术的迅速发展,传统的推荐系统已不能很好地适应基于位置的推荐服务,同时也面临隐私泄露的问题.本文针对上述问题,首先提出一种分布式隐私保护推荐框架,并利用差分隐私保护理论,设计基于分布式框架的奇异值分解推荐算法,同时利用保序加密函数实现用户请求位置的保护.理论分析和在两个真实的数据集上的实验表明,本文提出的...     

网络电视推荐系统框架及协同过滤算法的研究

针对网络电视推荐系统中通常采用的协同过滤推荐算法的不足,提出了一种将聚类、用户相似—信任关系和项目属性关系相组合的协同过滤推荐技术.该组合推荐技术首先通过聚类分析缩小用户的有效搜索范围,其次通过引入信任关系来提高推荐的准确性,从而为目标用户提供更好的推荐结果.经过实验表明,该算法提高了推荐质量.     

融入注意力网络的深度分解机推荐算法

推荐系统能够在海量的信息中找到满足用户个性化需求的信息。随着深度学习的发展,深度学习也开始广泛被推荐系统所应用。CTR预估在推荐系统中起着重要作用,已被应用在个性化推荐、信息检索、在线广告等多个领域。针对推荐系统数据量大且稀疏的问题,文中将注意力网络和xDeepFM模型融合,提出了一种新的基于深度学习的CTR预估模型,即Atte-xDeepFM模型。该模型能够解决特征稀疏问题,有效学习特征之间的交互关系,且不需要手动提取特征工程中的有用信息。在Avazu数据集和Criteo数据集上进行的对比实验证明了文中提出的模型的有效性。与推荐系统CTR预估常用的算法模型对比,文中所提出的模型具有更好的推荐效果。     

直播电视推荐系统的评分预测算法研究

伴随电视频道的不断增加,推荐系统在直播电视领域应用成为研究热点.然而,直播电视独特的播放和收视方式使得传统的VOD（Video On Demand）推荐系统无法直接应用,已有的推荐频道的方法不关注正在播出的节目状态从而影响了推荐准确率,而推荐节目的方法难以应对节目冷启动.为此,本文提出了一种融合频道推荐和节目推荐的评分预测算法OFAP（Over the First by Adding Preference）.首先,利用聚类方法对每个用户实现差异性的收视时段划分,构建他们的频道-时段偏好矩阵和预推荐评分权重矩阵;其次,提出一个评分替代策略使得已有的推荐节目的算法能够应对节目冷启动,从而实现预推荐;最后,通过融合用户偏好、预推荐评分权重与预推荐结果,构建评分预测函数,将预推荐算法的评分预测结果作为评分预测函数的训练样本.实验表明,采用Precision@N和Recall@N作为评价标准,本文所提方法OFAP明显优于对比算法.     

自动建立信任的防攻击推荐算法研究

 随着互联网中信息资源的日益增多,个性化推荐技术作为缓解"信息过载"的有效手段,得到了越来越多的研究者的关注.由于互联网天然的开放性,在商业利益的驱动下,部分恶意用户通过伪造虚假数据来影响系统的推荐结果,从而达到盈利的目的.本文提出一个自动建立信任的防攻击推荐算法,在考虑了用户评分相似性的基础上,引入适当的信任机制,通过为目标用户动态建立和维护有限数量的信任对象来获得可靠的推荐.大量基于真实数据集的实验表明,提出的算法能大大提高推荐系统的鲁棒性和可靠性,并在一定程度上提高了推荐的精准度.     

基于特征选择的推荐系统托攻击检测算法

基于协同过滤的电子商务推荐系统极易受到托攻击,托攻击者注入伪造的用户模型增加或减少目标对象的推荐频率,如何检测托攻击是目前推荐系统领域的热点研究课题.分析五种类型托攻击对不同协同过滤算法产生的危害性,提出一种特征选择算法,为不同类型托攻击选取有效的检测指标.基于选择出的指标,提出两种基于监督学习的托攻击检测算法,第一种算法基于朴素贝叶斯分类;第二种算法基于k近邻分类.最后,通过实验验证了特征选择算法的有效性,及两种算法的灵敏性和特效性.     

Social recommendation algorithm fusing user interest social network

Using the social information among users in recommender system can partly solve the data sparsely problems and significantly improve the performance of the recommendation system. However, the recommendation systems which using the users' social information have two main problems: the explicit user social connection information is not always available in real-world recommender systems, and the user social connection information is directly used in recommender systems when the user explicit social information is available. But as we know that the user social information is not all based on user interest, so this can introduce noise to the recommender systems. This paper proposes a social recommender system model called interest social recommendation (ISoRec). Based on probability matrix factorization (PMF), the model addresses the problems mentioned above by combining user-item rating matrix, explicit user social connection information and implicit user interest social connection information to make more accurately recommendation. In addition, the computational complexity of our algorithm is linear with respect to the number of observed data sets used in this algorithm, and can scalable to very large datasets.     

基于数据挖掘的个性化智能推荐系统应用研究

在当前家庭数字化日趋普及的环境下,为了给用户提供一种智能型、个性化的多媒体内容推荐服务,通过研究协同式信息过滤技术,结合数据挖掘技术,设计并实现了一个智能型、个性化的多媒体推荐系统。系统可以根据用户的使用习惯、使用时间、使用环境以及最近选择的项目进行分析,进行判断后列出最优推荐资源。系统通过研究个人信息的自我学习技术、个性化特征分析技术以及多媒体内容的搜寻技术,将上述技术应用在推荐服务系统中,具有一定的实际意义。     

Novelty‐driven recommendation by using integrated matrix factorization and temporal‐aware clustering optimization

With the rapid proliferation of information and communication technology (ICT), the vast amount of available data creates information overload. The Websites and e‐commerce applications employ several information filtering methods such as personalized recommender system to manage the information overload. The recommender system assists the users in obtaining the desired list of products based on their interest. Several existing research works focus on the novelty or unexpectedness in the recommendation list while ensuring the quality to enhance the recommendation mechanism. It is essential to balance the unexpected and useful products or services to generate the satisfactory personalized recommendations with novelty. Thus, this paper proposes a novelty‐driven movie suggestion using integrated matrix factorization and temporal‐aware clustering optimization (NOMINATE). The proposed approach determines the personalized preferences through probabilistic matrix factorization (PMF) and contextually updates the rules and extracts the user preferences based on the inherent features of both the users and movies with temporal information. The NOMINATE approach also suggests the novelty‐driven, and desired top‐N movies to the users through the K‐means, and particle swarm optimization (PSO)‐based clustering algorithm with the help of LOD source. To identify the expert users, the NOMINATE approach applies the K‐means and PSO‐based clustering algorithm to enrich the personalized features of the users. Moreover, it integrates the relevant features with the preferred set of features for each user using the LOD source and decides a set of optimal preferences of the users. Finally, the NOMINATE approach generates the top‐N recommendation list for the corresponding user through ranking method. The experiment results stipulate that the NOMINATE approach personalize the top‐N movie recommendations with high performance regarding accuracy and novelty when compared with the existing recommendation method.     

一种基于差分隐私和时序的推荐系统模型研究

推荐系统的建立依赖用户的个人隐私信息,攻击者可以通过推荐的结果对用户的状态和行为进行预测.目前,虽然有对基于协同过滤近邻隐私保护的研究,但是对基于模型的隐私保护的关注度并不够高.差分隐私理论定义了一个相当严格的防攻击模型,通过添加噪声使数据失真达到隐私保护的目的,而且用户的兴趣存在兴趣漂移问题,对推荐效果造成影响,因此,提出基于差分隐私理论和时序理论构建基于模型的推荐系统.首先,根据差分隐私理论,给用户的评分数据增加小波动的符合Laplace分布的噪声,增大待分解矩阵的安全系数;然后,在随机梯度下降模型的基础上,将时序因子建模为时间权重,提高模型的准确性.实验证明该算法的准确性,并且为增强隐私研究提供了新的思路.     

Performance improvement for recommender systems using ontology

Recommender systems have emerged in the e-commerce domain and have been developed to actively recommend appropriate items to online users. The use of recently developed hybrid recommendation systems has helped overcome the main drawbacks of Content-Based Filtering (CBF) and Collaborative Filtering (CF). In hybrid recommendation systems that combine CF and CBF, the CF part uses two methods, including memory- and model-based approaches. Both approaches have some advantages and disadvantages for item recommendation. Sparsity has been one of the main difficulties associated with these approaches, whereas recommendation with high accuracy has been one of the important advantages of the memory-based approach. However, this approach is not scalable for current recommendation systems as their databases include huge numbers of items and users. In contrast, the model-based approach generates recommendations with low accuracy but is scalable for large databases in e-commerce recommender systems. Accordingly, to address this problem and take advantage of both approaches, in this work, we propose a new hybrid recommendation method and evaluate it using a real-world dataset. The aim is to improve efficiency and accuracy by designing a heuristic hybrid recommender method that combines memory-based and model-based approaches. Specifically, we use ontology in the CF part and improve ontology structure by eliminating uniformity of edges of the hierarchical relation between concepts (IS-A relation) in item ontology in the CBF part. Ontology structure is considered for improving accuracy; according to this, a new method for measuring semantic similarity that is more accurate than the traditional methods is presented. This new method can enhance the accuracy of CF and CBF in our method. In addition, the number of searches required to find similar clusters and neighbor users of the target user is decreased significantly using ontology, enhanced clustering and the new proposed algorithm. We evaluate the proposed method using a real-world dataset. The experimental results show that our method is more scalable and accurate than the benchmark k-Nearest Neighbor (k-NN) and model-based recommendation methods.     

基于用户兴趣传播的协同过滤方法

推荐系统帮助用户过滤无用信息并预测其可能感兴趣的产品。在推荐系统中,协同过滤是应用最为广泛的方法之一。然而,传统的协同过滤方法是在产品维度上计算用户相似度,而且在计算相似度时无法考虑邻居用户的影响。因此,该类方法往往受到高维度、数据稀疏等问题的困扰。为此,本文提出一种基于用户兴趣传播的协同过滤方法,在兴趣维度上计算用户相似度,同时考虑了兴趣在不同用户间的传播。该方法不仅可以有效防止冷启动和数据稀疏问题,而且具有较高的预测准确度。在标准数据集MovieLens上的测试结果表明了本文算法的有效性。     

Optimisation of combined collaborative recommender systems

A new approach to collaborative user modelling is presented in this paper. We have developed a framework that can be used for easy testing of different concepts. We have also introduced three different areas where collaborative modelling can be further improved. For the first phase of testing, we have created a hybrid system based on three different collaborative recommender techniques. Since this system implements multiple collaboration techniques, we decided to call this approach Combined Collaborative Recommender. Although each prediction technique can produce adequate results, we have proved that the combination of these techniques into a unified system provides a much more stable system. It should also be pointed out that these analyses were done using a very large dataset (more than 2 million ratings) providing reliable results. Results of these optimisations are presented along with pointers for further development.     

Attribute Reduction with Rough Set in Context-Aware Collaborative Filtering

The problem of different contextual information to influence the user-item-context interactions at varying degrees in context-aware recommender systems is addressed.To improve the performance accuracy,we develop a novel attribute reduction algorithm in order to effectively extract the core contextual information using rough set.We combine collaborative filtering with contextual information significance to generate more accurate predictions.We experimentally evaluate our approach on UCI machine learning repository and two real world data sets.Experimental results demonstrate that our proposed approach outperforms existing state-of-theart context-aware recommendation methods.     

社会标签情感分析的个性化影视推荐算法研究

提出了一种在标签中引入情感分析的个性化推荐算法,该算法在两个共享资源的相似度计算中考虑了情感因子以改进已有的推荐算法。在实际数据集上比较了多种个性化推荐算法的命中率,实验结果证明了已提出算法的先进性。     

WWW collaborative recommendation based on reliability

Collaborative filtering recommender systems often suffer from the ＂Matchmaker＂ problem, which comes from the false assumption that users are counted only based on their similarity, and high similarity means good advisers. In order to find good advisers for every user, a matchmaker＇s reliability mode based on the algorithm deriving from Hits is constructed, and it is applied in the proposed World Wide Web （WWW） collaborative recommendation system. Comparative experimental results also show that our approach obviously improves the substantial performance.     

基于情景建模的移动互联网音乐推荐系统研究

近年在当下移动互联网中,数字音乐得到了大力研发,而基于大数据算法的数字音乐个性化推荐正在蓬勃发展,对于这类研究需要借助计算机技术、多媒体技术来补充.在探索过程中,线上音乐的传统二元推荐算法正逐渐被情景推荐所取代,通过构建情景模型的音乐推荐算法有利于用户获取音乐,可以增长互联网音乐的有效投递,通过传统音乐推荐算法与改进的...     

A survey of trust in internet applications

Trust is an important aspect of decision making for Internet applications and particularly influences the specification of security policy, i.e., who is authorized to perform actions as well as the techniques needed to manage and implement security to and for the applications. This survey examines the various definitions of trust in the literature and provides a working definition of trust for Internet applications. The properties of trust relationships are explained and classes of different types of trust identified in the literature are discussed with examples. Some influential examples of trust management systems are described.     

Transformation of a health care information system: aself-assessment survey

US health care providers are seeking to reduce costs while simultaneously maintaining quality of care. One strategy for the reinvention of the health care industry is the more effective use of new quality and information technology solutions. Results are presented from a survey of 98 top executives at Baylor Health Care System (BHCS), a large, multifunction health care organization in Dallas, TX. The survey sought to assess the executives' perceptions of current BHCS quality practices. The study used a survey developed for the health care industry based on the Malcom Baldrige National Quality Award (MBNQA) criteria. This paper reports findings related to the quality of BHCS internal and external data and information quality. Factor analysis and regression models using the survey data were used to highlight findings that include: (1) the need for a $50+ million information system transformation at BHCS as an essential action item to achieve the organization's critical success factors, and (2) the importance of internal and external data and information in achieving business process redesign and a quality transformation at BHCS. Results highlight the need for further research investigating the dimensions associated with the MBNQA criteria and their relationship with the information and analysis component