面向群组用户时序行为的动态推荐算法

针对现实系统中用户偏好随时间动态变化且一个用户ID背后可能是一个家庭的多个成员在共用的问题,提出一种为这类隐含多个类型成员行为的群组用户解决其偏好随时间而变化的动态推荐算法。首先,假设用户的历史行为数据包括曝光数据和点击数据,并通过学习当前时刻下群组用户的各类型角色权重来判别当前成员角色;其次,根据曝光数据提出两种设计思路来构造流行度模型,并采用逆倾向评分加权方法来平衡训练数据;最后,利用矩阵分解技术得出随时间变化的用户潜在偏好因子和物品潜在属性因子,计算两者内积后得出用户随时间变化的Top-K偏好推荐。实验结果表明,该算法在召回率、平均精度均值（MAP）、归一化折损累计增益（NDCG）这三个指标上一天24个时刻中均能有至少16个时刻的表现优于基准方法,并能缩短运行时间,降低计算的时间复杂度     

基于张量相似度的推荐方法研究

传统推荐方法中将用户建模为向量的建模方式只关注用户单方面偏好，为了弥补此种建模方法的局限性，提出一种将用户建模为矩形的张量建模方法。构建了一个基于融合协同过滤与序列推荐算法的推荐模型，该模型集成了Fastformer模型和键值记忆网络对用户张量进行建模；结合用户张量与目标物品的距离及偏置项对用户张量与目标物品的相似度进行计算。在MovieLens和CiaoDVD数据集上对该模型进行实验验证，实验结果表明，该模型能够关注用户多方面偏好并在推荐结果的精准度上优于基线方法，特别是在HR与NDCG评价指标上分别比现有基线方法平均提高了1.4%、1.95%。     

引入兴趣稳定性的时间敏感协同过滤算法

针对物品流行偏置现象,将物品流行度引入到用户兴趣中建模,提出了基于物品流行度的用户兴趣特征相似度模型。针对传统模型没有考虑到用户兴趣稳定性和难以实时捕获用户兴趣问题,在计算用户兴趣相似度过程中引入时间敏感和兴趣稳定性,提出了引入时间敏感的用户兴趣稳定性的相似度模型。最后,融合权重因子将两种相似度模型线性加权,提出了一种新颖的IPSTS算法模型。实验表明,该算法相比传统协同过滤算法在平均绝对误差（MAE）和均方根误差（RMSE）上均有明显降低的同时也能挖掘出长尾物品,缓解了物品流行偏置现象。     

基于时空信息和任务流行度分析的移动群智感知任务推荐

现有移动群智感知任务推荐的共同缺点是：一方面,未充分考虑时空信息对工人偏好的影响,导致推荐准确性低;另一方面,忽略了任务流行度对推荐的影响,导致推荐覆盖率差。为解决这些问题,提出一种基于时空信息和任务流行度分析的移动群智感知任务推荐方法。充分利用工人执行记录中的相关信息（如工人执行任务的时间、位置）,准确预测工人对任务的偏好;基于工人声誉和任务执行情况分析任务流行度并设计任务流行度惩罚因子,提升推荐效果的覆盖率;结合工人偏好和流行度惩罚因子生成任务推荐列表。实验结果表明,与现有基线方法相比,所提出方法在推荐准确率上平均提升了3.5%,推荐覆盖率上平均提高了25%。     

融合多元信任关系的新型推荐模型

现有的基于信任关系的推荐模型大多用于预测缺失的“用户-物品”评分,未考虑信任关系自身的多样性。基于此,研究了考虑多元信任关系的物品序列推荐的矩阵分解模型。首先,针对社交网络中目标用户可分别作为信任者和受信任者2种情形,提出相应的信任者相似度因子模型(truster factored similarity model,TrusterFSM)和受信任者相似度因子模型(trustee factored similarity model, TrusteeFSM);然后,通过融合两者获得最终的信任相似度因子模型(trust factored similarity model, TrustFSM),TrustFSM引入了物品偏置、用户偏置、物品相似度、用户相似度和社交相似度的影响。最后,基于真实数据集的实验结果表明,TrustFSM获得了最优的推荐效果,所提3种模型的性能优于现有模型（有/无引入信任者信息）。     

基于阶段时序效应的奇异值分解推荐模型

针对传统基于时序效应的奇异值分解（SVD）推荐模型在对用户预测评分建模过程中只考虑评分矩阵,采用复杂的时间函数拟合项目的生命周期、用户偏好的时序变化过程,造成模型难于解释、用户偏好捕获不准、评分预测精度不够高等问题,提出了一种改进的综合考虑评分矩阵、项目属性、用户评论标签和时序效应的推荐模型。首先,通过将时间轴划分时间段,利用sigmoid函数将项目的阶段流行度变换为[0,1]区间上的影响力来改进项目偏置;其次,利用非线性函数将用户偏置的时序变化转变为阶段评分均值与总体均值偏差的时序变化来改进用户偏置;最后,通过捕获用户对项目的阶段兴趣度,结合其相似用户在此时间段对该项目的好评率,生成用户项目交互作用影响因子,实现用户项目交互作用的改进。在Movielence 10M和20M电影评分数据集上的测试表明,改进模型能更好地捕获用户偏好的时序变化过程,提高评分预测准确性,均方根误差平均提高了2.5%。     

基于改进型启发式相似度模型的协同过滤推荐方法

为提高协同过滤推荐方法的准确性和有效性,提出一种基于改进型启发式相似度模型的协同过滤推荐方法PSJ。该方法考虑了用户评分差值、用户全局评分偏好和用户共同评分物品数三个因素。PSJ方法的Proximity因子使用指数函数反映用户评分差值对用户相似度的影响,这样也可避免零除问题;将NHSM方法中的Significance因子和URP因子合并成PSJ方法的Significance因子,这使得PSJ方法的计算复杂度低于NHSM方法;而且为了提高在数据稀疏情况下的推荐效果,PSJ方法同时考虑了用户间的评分差值和用户全局评分两个因素。实验采用Top-k推荐中的查准率和查全率作为衡量标准。实验结果表明,当推荐物品数大于20时,与NHSM、杰卡尔德算法、自适应余弦相似度（ACOS）算法、杰卡尔德均方差（JMSD）算法和皮尔逊相关系数算法（SPCC）相比,PSJ方法的查准率与查全率均有提升。     

基于情感分析和LDA主题模型的协同过滤推荐算法

协同过滤推荐算法通常基于物品或用户的相似度来实现个性化推荐,但是数据的稀疏性往往导致推荐精度不理想。大多数传统推荐算法仅考虑用户对物品的总体评分,而忽略了评论文本中用户对物品各个属性面的偏好。该文提出一种基于情感分析的推荐算法SACF(reviews sentiment analysis for collaborative filtering),该算法在经典的协同过滤推荐算法的基础上,考虑评论文本对相似度计算的影响。SACF算法利用LDA主题模型挖掘物品潜在的K个属性面,通过用户在各个属性面上的情感偏好计算用户相似度,从而构建推荐模型。基于京东网上评论数据集的实验结果表明,SACF算法不但可以有效地改善传统协同过滤推荐算法中数据稀疏性的问题,而且提高了推荐系统的精度。     

融合显隐式反馈的协同过滤推荐模型

融合显式和隐式反馈已被应用于提升推荐模型的性能,但是,现有的此类推荐模型未能保留显式反馈中反映用户偏好程度的信息,且现有研究认为拥有显式反馈的数据和仅拥有隐式反馈的数据对于模型具有同等影响,未能充分发挥显式反馈的优势.针对这些问题,提出一种新的融合显式和隐式反馈的协同过滤推荐模型(CEICF).首先,所提出模型提取显式反馈中的特征得到用户/物品的全局偏好向量;然后,从隐式反馈中提取用户/物品的潜在向量,进而将两种向量进行融合得到用户/物品的偏好向量;最后,使用神经网络预测用户与物品交互的可能性.在训练模型时,定义一种加权的二进制交叉熵损失函数,加强显式反馈对模型的影响来增强模型捕获用户偏好的能力.为了验证所提出模型的有效性,在覆盖不同领域的现实数据集上进行实验,实验结果表明,CEICF可有效地融合显式和隐式反馈,且推荐效果相对于基线模型有显著提升.     

融合全局信息的注意力增强会话推荐方法

近年来,基于会话推荐系统（session-based recommender system,SRS）的应用和研究是推荐系统的一个热门方向。如何利用用户会话信息进一步提升用户满意度和推荐精确度,是基于会话推荐系统的主要任务。目前大多数SBR模型仅基于目标会话对用户偏好建模,忽略了来自其他会话的物品转换信息,导致无法全面了解用户偏好。为了解决其局限性,提出融合全局上下文信息注意力增强的图神经网络模型（global context information graph neural networks for session-based recommendation,GCI-GNN）。该模型利用所有会话上的物品转换关系,更准确地获取用户偏好。具体而言,GCI-GNN从目标会话和全局会话学习物品向量表示。使用位置感知注意网络,将反向位置信息纳入物品嵌入中。考虑会话长度信息学习用户表示进而达到更有效的推荐。在Diginetica和Yoochoose数据集上进行实验,实验结果表明,相对最优的基准模型,GCI-GNN模型在Diginetica数据集各项指标上的提高超过2个百分点,在Yoochoose数据集上,GCI-GNN模型在各项指标上的提高超过1个百分点,验证了GCI-GNN模型的有效性。     

融合多模态深度游走与偏差校准因子的推荐模型

曝光偏差严重影响协同过滤模型的推荐精度,导致预测结果偏离用户的真实兴趣,而现有模型对曝光偏差的建模能力有限,甚至放大偏差。为此,提出融合多模态深度游走与偏差校准因子（MmDW-BC）的推荐模型。首先,引入项目多模态属性特征作为项目图的连接边,从而缓解低曝光项目交互数据稀疏的问题;在此基础上,构建图嵌入模块——多模态深度游走（MmDW）将项目多模态信息融入嵌入向量,以获取丰富的节点表示;最后,基于校准策略设计新的偏差校准推荐算法进行用户偏好预测。将提出的模型应用于Amazon和ML-1M数据集上,实验结果验证所提模型明确考虑曝光偏差来提升推荐精度的必要性和有效性。     

基于卷积神经网络交互的用户属性偏好建模的推荐模型

潜在因子模型（LFM）以其优异的性能在推荐领域得到了广泛应用。在LFM中除了使用交互数据以外,辅助信息也被引入用于解决数据稀疏的问题,从而提升推荐的性能。然而,大多数LFM仍然存在一些问题：第一,LFM在对用户进行建模时,忽略了用户如何根据其特征偏好对项目作出决策;第二,采用内积的特征交互假设特征维度之间是相互独立的,而没有考虑到特征维度之间的关联。针对上述问题,提出一种新的推荐模型：基于卷积神经网络（CNN）交互的用户属性偏好建模的推荐模型（UAMC）。该模型首先获得用户的一般偏好、用户属性和项目嵌入,然后将用户属性和项目嵌入进行交互,以探索用户不同的属性对不同项目的偏好;接着将交互过的用户偏好属性送入CNN层来探索不同偏好属性的不同维度的关联,从而得到用户的属性偏好向量;接着使用注意力机制结合用户的一般偏好和CNN层得到的属性偏好,从而获得用户的向量表示;最后采用点积来计算用户对项目的评分。在Movielens-100K、Movielens-1M和Book-crossing这三个真实的数据集上进行了实验。实验结果表明,所提模型在均方根误差（RMSE）上与稀疏数据预测的神经网络分解机（NFM）模型相比分别降低了1.75%、2.78%和0.25%,验证了在LFM的评分预测推荐中,UAMC在提升推荐精度上的有效性。     

Collaborative user modeling with user-generated tags for social recommender systems

With the popularity of social media services, the sheer amount of content is increasing exponentially on the Social Web that leads to attract considerable attention to recommender systems. Recommender systems provide users with recommendations of items suited to their needs. To provide proper recommendations to users, recommender systems require an accurate user model that can reflect a user’s characteristics, preferences and needs. In this study, by leveraging user-generated tags as preference indicators, we propose a new collaborative approach to user modeling that can be exploited to recommender systems. Our approach first discovers relevant and irrelevant topics for users, and then enriches an individual user model with collaboration from other similar users. In order to evaluate the performance of our model, we compare experimental results with a user model based on collaborative filtering approaches and a vector space model. The experimental results have shown the proposed model provides a better representation in user interests and achieves better recommendation results in terms of accuracy and ranking.     

Collaborative filtering and deep learning based recommendation system for cold start items

Recommender system is a specific type of intelligent systems, which exploits historical user ratings on items and/or auxiliary information to make recommendations on items to the users. It plays a critical role in a wide range of online shopping, e-commercial services and social networking applications. Collaborative filtering (CF) is the most popular approaches used for recommender systems, but it suffers from complete cold start (CCS) problem where no rating record are available and incomplete cold start (ICS) problem where only a small number of rating records are available for some new items or users in the system. In this paper, we propose two recommendation models to solve the CCS and ICS problems for new items, which are based on a framework of tightly coupled CF approach and deep learning neural network. A specific deep neural network SADE is used to extract the content features of the items. The state of the art CF model, timeSVD++, which models and utilizes temporal dynamics of user preferences and item features, is modified to take the content features into prediction of ratings for cold start items. Extensive experiments on a large Netflix rating dataset of movies are performed, which show that our proposed recommendation models largely outperform the baseline models for rating prediction of cold start items. The two proposed recommendation models are also evaluated and compared on ICS items, and a flexible scheme of model retraining and switching is proposed to deal with the transition of items from cold start to non-cold start status. The experiment results on Netflix movie recommendation show the tight coupling of CF approach and deep learning neural network is feasible and very effective for cold start item recommendation. The design is general and can be applied to many other recommender systems for online shopping and social networking applications. The solution of cold start item problem can largely improve user experience and trust of recommender systems, and effectively promote cold start items.     

基于会话的多粒度图神经网络推荐模型

基于会话的推荐旨在根据当前用户的匿名会话的点击序列信息来预测用户的下一次点击行为。现有方法多数都是通过对用户会话点击序列的物品信息进行建模,并学习物品的向量表示,进而进行推荐。而作为一种粗粒度的信息,物品的类别信息对物品有聚合作用,可作为物品信息的重要补充。基于此,提出了基于会话的多粒度图神经网络推荐模型（SRMGNN）。首先,使用图神经网络（GNN）得到会话序列中的物品和物品类别的嵌入向量表示,并使用注意力网络捕捉用户的注意力信息;然后,将赋予了不同注意力权重值的物品和物品类别信息进行融合后,输入到门限循环单元（GRU）里;最后,通过GRU学习会话序列的物品时序信息,并给出推荐列表。在公开的Yoochoose数据集和Diginetica数据集上进行实验,实验结果验证了该模型在增加了物品类别信息后的优势,且实验结果表明了在Precision@20和MRR@20这2种评价指标上,该模型相较于短期注意力/记忆优先级（STAMP）模型、神经注意力（NARM）模型、GRU4REC等8种模型均有更好的效果。     

结合注意力机制的知识感知推荐算法

知识图谱在推荐系统中的应用越来越受重视,可以有效地解决推荐系统中存在的数据稀疏性和冷启动问题.但现有的基于路径和基于嵌入的知识感知推荐算法在合并知识图谱中的实体来表示用户时,并没有考虑到实体对于用户的重要性并不相同,推荐结果会受到无关实体的影响.针对现有方法的局限性,提出了一种新的结合注意力机制的知识感知推荐算法,并给...     

一种解决新项目冷启动问题的推荐算法

推荐系统作为缓解信息过载问题的有效方法之一,在社交媒体中的作用日趋重要.但是,新项目冷启动和新用户冷启动问题是推荐技术面临的难题.为了解决新项目冷启动问题,提出了用户时间权重信息概念,该定义考虑到了用户评价时间与项目发布时间的时间间隔,根据用户时间权重值的大小,可以判断该用户是积极用户还是消极用户,以及用户对新项目的偏爱程度;利用三分图的形式来描述用户-项目-标签、用户-项目-属性之间的关系.在充分考虑用户、标签、项目属性、时间等信息基础上,获得个性化的预测评分值公式,提出了推荐算法.实验结果表明:所提出的方法能够实现满足不同用户、不同偏好的个性化推荐,在为用户推荐到合适项目的同时还能带来惊喜.比较实验说明,所提出的方法推荐准确度高,推荐新颖度高.交叉验证实验结果表明:该方法在解决推荐算法中的新项目冷启动问题上,无论是在推荐的准确度还是推荐项目的新颖度上都是有效的.     

Modeling and broadening temporal user interest in personalized news recommendation

User profiling is an important step for solving the problem of personalized news recommendation. Traditional user profiling techniques often construct profiles of users based on static historical data accessed by users. However, due to the frequent updating of news repository, it is possible that a user’s fine-grained reading preference would evolve over time while his/her long-term interest remains stable. Therefore, it is imperative to reason on such preference evaluation for user profiling in news recommenders. Besides, in content-based news recommenders, a user’s preference tends to be stable due to the mechanism of selecting similar content-wise news articles with respect to the user’s profile. To activate users’ reading motivations, a successful recommender needs to introduce “somewhat novel” articles to users.In this paper, we initially provide an experimental study on the evolution of user interests in real-world news recommender systems, and then propose a novel recommendation approach, in which the long-term and short-term reading preferences of users are seamlessly integrated when recommending news items. Given a hierarchy of newly-published news articles, news groups that a user might prefer are differentiated using the long-term profile, and then in each selected news group, a list of news items are chosen as the recommended candidates based on the short-term user profile. We further propose to select news items from the user–item affinity graph using absorbing random walk model to increase the diversity of the recommended news list. Extensive empirical experiments on a collection of news data obtained from various popular news websites demonstrate the effectiveness of our method.     

Sequence-based trust in collaborative filtering for document recommendation

Collaborative filtering (CF) recommender systems have emerged in various applications to support item recommendation, which solve the information-overload problem by suggesting items of interest to users. Recently, trust-based recommender systems have incorporated the trustworthiness of users into CF techniques to improve the quality of recommendation. They propose trust computation models to derive the trust values based on users' past ratings on items. A user is more trustworthy if s/he has contributed more accurate predictions than other users. Nevertheless, conventional trust-based CF methods do not address the issue of deriving the trust values based on users' various information needs on items over time. In knowledge-intensive environments, users usually have various information needs in accessing required documents over time, which forms a sequence of documents ordered according to their access time. We propose a sequence-based trust model to derive the trust values based on users' sequences of ratings on documents. The model considers two factors – time factor and document similarity – in computing the trustworthiness of users. The proposed model enhanced with the similarity of user profiles is incorporated into a standard collaborative filtering method to discover trustworthy neighbors for making predictions. The experiment result shows that the proposed model can improve the prediction accuracy of CF method in comparison with other trust-based recommender systems.     

Deep collaborative filtering with social promoter score-based user-item interaction: a new perspective in recommendation

Most of the existing recommender systems understand the preference level of users based on user-item interaction ratings. Rating-based recommendation systems mostly ignore negative users/reviewers (who give poor ratings). There are two types of negative users. Some negative users give negative or poor ratings randomly, and some negative users give ratings according to the quality of items. Some negative users, who give ratings according to the quality of items, are known as reliable negative users, and they are crucial for a better recommendation. Similar characteristics are also applicable to positive users. From a poor reflection of a user to a specific item, the existing recommender systems presume that this item is not in the user’s preferred category. That may not always be correct. We should investigate whether the item is not in the user’s preferred category, whether the user is dissatisfied with the quality of a favorite item or whether the user gives ratings randomly/casually. To overcome this problem, we propose a Social Promoter Score (SPS)-based recommendation. We construct two user-item interaction matrices with users’ explicit SPS value and users’ view activities as implicit feedback. With these matrices as inputs, our attention layer-based deep neural model deepCF_SPS learns a common low-dimensional space to present the features of users and items and understands the way users rate items. Extensive experiments on online review datasets present that our method can be remarkably futuristic compared to some popular baselines. The empirical evidence from the experimental results shows that our model is the best in terms of scalability and runtime over the baselines.