Efficient and non-parametric reasoning over user preferences

We consider the problem of modeling and reasoning about statements of ordinal preferences expressed by a user, such as monadic statement like “X is good,” dyadic statements like “X is better than Y,” etc. Such qualitative statements may be explicitly expressed by the user, or may be inferred from observable user behavior. This paper presents a novel technique for efficient reasoning about sets of such preference statements in a semantically rigorous manner. Specifically, we propose a novel approach for generating an ordinal utility function from a set of qualitative preference statements, drawing upon techniques from knowledge representation and machine learning. We provide theoretical evidence that the new method provides an efficient and expressive tool for reasoning about ordinal user preferences. Empirical results further confirm that the new method is effective on real-world data, making it promising for a wide spectrum of applications that require modeling and reasoning about user preferences.     

Recommendation algorithm based on item quality and user rating preferences

Recommender systems are one of the most important technologies in e-commerce to help users filter out the overload of information. However, current mainstream recommendation algorithms, such as the collaborative filtering CF family, have problems such as scalability and sparseness. These problems hinder further developments of recommender systems. We propose a new recommendation algorithm based on item quality and user rating preferences, which can significantly decrease the computing complexity. Besides, it is interpretable and works better when the data is sparse. Through extensive experiments on three benchmark data sets, we show that our algorithm achieves higher accuracy in rating prediction compared with the traditional approaches. Furthermore, the results also demonstrate that the problem of rating prediction depends strongly on item quality and user rating preferences, thus opens new paths for further study.     

考虑用户偏好的启发式动态共乘匹配算法

针对现有算法很少考虑用户之间的共乘偏好需求,提出了一种考虑用户偏好的启发式动态共乘匹配算法。构建一个满足用户偏好需求的动态共乘匹配模型,旨在最大化系统匹配率和最小化车辆的绕行距离。算法首先根据出行请求的时间约束、车辆与用户的出行轨迹以及用户的兴趣偏好,过滤不满足用户偏好需求的车辆;其次,构建一个临时匹配图,设置边的权值为出行请求插入到车辆的当前行驶路线中的最小绕行距离;最后采用贪婪方式实现用户与车辆之间的匹配,并采用节点插入方式,将出行请求的出发地点和到达地点插入到车辆的当前行驶路线中。仿真结果表明,提出的启发式动态共乘匹配算法使车辆增加的平均绕行距离和运行时间低于现有算法,系统匹配率高于现有算法;用户的出行时间需求、兴趣偏好、信誉度等共乘需求对系统匹配率有显著影响。     

基于用户偏好的协作内容缓存策略

随着无线网络不断增长的业务需求,蜂窝架构频谱资源受限,回程容量将成为系统瓶颈。为了缓解这种瓶颈,考虑一种特殊的异构蜂窝网络,结合缓存节点的部署、用户位置分布、用户对请求内容的偏好以及缓存节点有限的存储空间,对内容存储及用户关联联合优化问题进行建模分析。将目标函数建模为请求时延的最小化,简单证明该问题是NP-hard的,并设计了基于改进KM(Kuhn-Munkres)的内容放置策略。最后,通过实验比较了该算法与其他基准方案的性能。     

基于评论挖掘和用户偏好学习的评分预测协同过滤*

提出了一种新的方法挖掘评论中的文字信息,将评论对象被用户关注的层面发掘出来并评分,根据这些层面的分数以及用户过往的评分数据学习出用户的偏好,最后根据用户的偏好预测其他待评分对象的分数并产生推荐。实验结果表明,提出的方法在预测准确度方面较传统方法有一定程度的提高。     

结合信任机制和用户偏好的协同过滤推荐算法

仅凭相似度来定位邻居用户对传统协同过滤算法的性能有严重的负面影响。引入社会网络中的信任机制,从个体在社交圈中的主观信任和全局声誉角度出发建模。分别考虑用户交互、评分差和用户偏好调节生成直接信任度。利用声誉及专家信任优先模型聚合生成间接信任度,将两者动态加权形成用户之间的信任关系。用参数[η]协调信任和相似双属性,使用户关系更加紧密,有效地解决新用户和稀疏性问题。经实证,改良后的模型颇有成效。     

一种在用户偏好不确定情况下的Web服务选择方法*

首先由用户对各个QoS属性的偏好给出语言描述及其不确定度;然后通过查找对照表将其换算成各个QoS属性的权重系数;最后使用QoS属性值和权重系数进行候选服务的综合评价,得到最接近满足用户不确定偏好的候选服务。模拟实验结果证明了该方法的有效性。     

基于用户偏好的信任网络随机游走推荐模型

为了提高推荐算法评分预测的准确度,解决冷启动用户推荐问题,在TrustWalker模型基础上提出一种基于用户偏好的随机游走模型——PtTrustWalker。首先,利用矩阵分解法对社会网络中的用户、项目相似度进行计算;其次,将项目进行聚类,通过用户评分计算用户对项目类的偏好和不同项目类下的用户相似度;最后,利用权威度和用户偏好将信任细化为不同类别下用户的信任,并在游走过程中利用信任用户最高偏好类中与目标物品相似的项目评分进行评分预测。该模型降低了噪声数据的影响,从而提高了推荐结果的稳定性。实验结果表明,PtTrustWalker模型在推荐质量和推荐速度方面相比现有随机游走模型有所提高。     

动态信任感知的偏好演化模型研究

有界置信模型（bounded confidence model,BCM）是舆论动力学中对偏好演化进行建模的重要模型,但其假设个体会完全接受与之交流个体的偏好以及所有的个体会诚实表达其偏好与实际情况不符,存在明显不足。针对该问题,提出了一种动态信任感知的偏好演化模型（dynamic trust-aware preferences evolution model,DTPEM）。首先,引入偏好接受度算子建模个体对交互对象偏好的接受度;其次,引入动态信任度算子度量个体偏好表达的诚实程度;然后建模偏好差距对信任度的影响。基于智能体模拟仿真实验与其他模型进行对比,结果表明,DTPEM在偏好演化的准确度上有了较大幅度的提升。     

融合用户特征优化聚类的协同过滤算法

针对推荐系统领域中应用最广泛的协同过滤推荐算法仍伴随着数据稀疏性、冷启动和扩展性问题,基于用户冷启动和扩展性问题,提出了基于改进聚类的PCEDS(pearson correlation coefficient and euclidean distance similarity)协同过滤推荐算法。首先针对用户属性特征,采用优化的K-means聚类算法对其聚类,然后结合基于信任度的用户属性特征相似度模型和用户偏好相似度模型,形成一种新颖的PCEDS相似度模型,对聚类结果建立预测模型。实验结果表明:提出的PCEDS算法比传统的协同过滤推荐算法在均方根误差(RMSE)上降低5%左右,并且推荐准确率(precision)和召回率(recall)均有明显提高,缓解了冷启动问题,同时聚类技术可以节省系统内存计算空间,从而提高了推荐效率。     

一种上下文移动用户偏好自适应学习方法

针对移动网络对个性化移动网络服务系统的性能提出了更高的要求,但现有研究难以自适应地修改上下文移动用户偏好以为移动用户提供实时、准确的个性化移动网络服务的问题,提出了一种上下文移动用户偏好自适应学习方法,在保证精确度的基础上缩短了学习的响应时间.首先,通过分析移动用户行为日志来判断移动用户行为是否受上下文影响,并在此基础上判断移动用户行为是否发生变化.然后,根据判断结果对上下文移动用户偏好进行修正.在对发生变化的上下文移动用户偏好进行学习时,将上下文引入到最小二乘支持向量机中,进一步提出了基于上下文最小二乘支持向量机(C-LSSVM)的上下文移动用户偏好学习方法.最后,实验结果表明,当综合考虑精确度和响应时间两方面因素时,所提出的方法优于其他学习方法,并且可应用于个性化移动网络服务系统中.     

The influence of affordances on user preferences for multimedia language learning applications

This study investigates the influence of sensory and cognitive affordances on the user experience of mobile devices for multimedia language learning applications. A primarily audio-based language learning application – ‘Vowel Trainer’, was chosen against a comparison, text and picture-based language learning application – ‘Learn English for Taxi Drivers’. Impressions of the two applications were assessed on two different devices that have virtually the same interface and identical sound output (when headphones are used), but differ in physical size: the iPhone and the iPad. A mixed design was chosen, with native language as a group factor and device type (iPad vs. iPhone) and language application type (audio vs. video) as within groups factors. Assessments of sensory and cognitive affordances were made, along with measurement of learner preferences of each application. Data from 41 participants (21 native English speakers, 20 non-native English speakers) were analysed, revealing device differences in both audio and visual subjective quality ratings, despite only visual quality being affected by the device's physical limitations. We suggest that sensory affordances (indexed by subjective quality) are not simply a function of physical limitations, but are heavily influenced by context. The implications for developing design guidelines for language learning and other multimedia applications are discussed.     

融入偏好的区间高维多目标集合进化优化方法

尽管区间参数高维多目标优化问题普遍存在且非常重要, 但是, 目前求解该问题的方法却很少. 本文提出一种有效解决该问题的集合进化优化方法, 通过在进化过程中融入决策者的偏好, 以得到符合决策者偏好的Pareto解集. 该方法将原优化问题转化为以超体积、不确定度、决策者满意度为新目标的确定型3目标优化问题; 为了求解转化后的优化问题, 采用集合Pareto占优关系比较个体, 并设计融入决策者偏好的延展性测度, 以进一步区分具有相同序值的个体; 此外, 还提出集合变异与重组策略, 以生成高性能的子代种群. 采用4个基准高维多目标优化问题和1个汽车驾驶室设计问题测试所提方法的性能, 并将其与另外3种方法进行对比. 实验结果验证, 该方法能得到收敛性、延展性、不确定度, 以及决策者满意度均衡的Pareto解集.     

Predicting user preferences of environment design: a perceptual mechanism of user interface customisation

It is a well-known fact that users vary in their preferences and needs. Therefore, it is very crucial to provide the customisation or personalisation for users in certain usage conditions that are more associated with their preferences. With the current limitation in adopting perceptual processing into user interface personalisation, we introduced the possibility of inferring interface design preferences from the user’s eye-movement behaviour. We firstly captured the user’s preferences of graphic design elements using an eye-tracker. Then we diagnosed these preferences towards the region of interests to build a prediction model for interface customisation. The prediction models from eye-movement behaviour showed a high potential for predicting users’ preferences of interface design based on the paralleled relation between their fixation and saccadic movement. This mechanism provides a novel way of user interface design customisation and opens the door for new research in the areas of human–computer interaction and decision-making.     

Recommendation algorithm based on item quality and user rating preferences

Recommender systems are one of the most im- portant technologies in e-commerce to help users filter out the overload of information. However, current mainstream recommendation algorithms, such as the collaborative filter- ing CF family, have problems ness. These problems hinder such as scalability and sparse- further developments of rec- ommender systems. We propose a new recommendation al- gorithm based on item quality and user rating preferences, which can significantly decrease the computing complexity. Besides, it is interpretable and works better when the data is sparse. Through extensive experiments on three benchmark data sets, we show that our algorithm achieves higher accu- racy in rating prediction compared with the traditional ap- proaches. Furthermore, the results also demonstrate that the problem of rating prediction depends strongly on item quality and user rating preferences, thus opens new paths for further study.     

Modelling user preferences and mediating agents in electronic commerce

An important ingredient in agent-mediated electronic commerce is the presence of intelligent mediating agents that assist electronic commerce participants (e.g. individual users, other agents, organisations). These mediating agents are in principle autonomous agents that interact with their environments (e.g. other agents and web-servers) on behalf of participants who have delegated tasks to them. For mediating agents a (preference) model of participants is indispensable. In this paper, a generic mediating agent architecture is introduced. Furthermore, we discuss our view of user preference modelling and its need in agent-mediated electronic commerce. We survey the state of the art in the field of preference modelling and suggest that the preferences of electronic commerce participants can be modelled by learning from their behaviour. In particular, we employ an existing machine learning method called inductive logic programming (ILP). We argue that this method can be used by mediating agents to detect regularities in the behaviour of the involved participants and induce hypotheses about their preferences automatically. Finally, we discuss some advantages and disadvantages of using inductive logic programming as a method for learning user preferences and compare this method with other approaches.     

A dempster-shafer approach to modeling agent preferences for plan recognition

Plan recognition is an important task whenever a system has to take into account an agent's actions and goals in order to be able to react adequately. Most plan recognizers work by merely maintaining a set of equally plausible plan hypotheses each of which proved compatible with recent observations without taking into account individual preferences of the currently observed agent. Such additional information provides a basis for ranking the hypotheses so that the best one can be selected whenever the system is forced to react (e.g., to provide help to the user of a software system to accomplish his goals). Furthermore, hypotheses with low valuations can be excluded from considerations at an early stage. In this paper, an approach to the quantitative modeling of the required agent-related data and their use in plan recognition is presented. It relies on the DempsterShafer Theory and provides mechanisms for the initialization and update of corresponding numerical values.     

基于改进用户属性评分的协同过滤算法

为解决在基于用户的推荐算法中,用户相似度计算精度较低、缺乏个性化等问题,提出一种基于改进用户属性评分的协同过滤算法(IUAS-CF)。针对个性用户、偏执用户等在评分矩阵上存在的评价值范围差异,基于现有的相似度计算公式设计一种适应于计算个性化用户相似度的距离度量公式;针对用户自身存在影响用户抉择的用户属性,设法将用户属性评分量化,将其引入相似度计算公式中。实验结果表明,IUAS-CF算法能更真实地反映用户评分偏好,提高了推荐系统的推荐精度,更好地满足了用户对系统的个性化需求。     

融合协同过滤与用户偏好的旅游组推荐方法

近年来，组推荐系统已经逐渐成为旅游推荐领域的研究热点之一。传统的推荐系统面临的数据稀疏性问题在组推荐系统中同样存在。基于评分的推荐系统中，可以把组推荐系统分为对单个用户的偏好预测和对组内成员预测结果的融合两个阶段。为提高推荐的效果，提出了一种融合协同过滤与用户偏好的旅游组推荐方法，它考虑了用户的预测评分和组推荐结果的准确性。在协同过滤中通过加入相似性影响因子和关联性因子进行预测评分，然后在均值策略和最小痛苦策略的基础上，提出了满意度平衡策略，该策略考虑了组内成员的局部满意度和整体满意度。实验表明，所提出的方法提高了推荐的准确率。     

Effects of smartphone icon background shapes and figure/background area ratios on visual search performance and user preferences

Smartphones are becoming increasingly popular, users are provided with various interface styles with different designed icons. Icon, as an important competent of user interface, is regarded to be more efficient and pleasurable. However, compared with desktop computers, fewer design principles on smartphone icon were proposed. This paper investigated the effects of icon background shape and the figure/background area ratio on visual search performance and user preference. Icon figures combined with six different geometric background shapes and five different figure/background area ratios were studied on three different screens in experiments with 40 subjects. The results of an analysis of variance (ANOVA) showed that these two independent variables (background shape and figure/background area ratio) significantly affected the visual search performance and user preference. On 3.5-in (1 in=0.025 4 m) and 4.0-in displays, unified backgroundwould be optimal, shapes such as square, circle and transitions between them (e.g., rounded square, squircle, etc.) are recommended because backgrounds in these shapes yield a better search time performance and subjective satisfaction for ease of use, search and visual preference. A 60% figure/background area ratio is the most appropriate for smartphone icon design on the 3.5-in screen, while a 50% area ratio could be a suggestion for both relatively optimized search performance and user preference on 4.0-in. In terms of the 4.7-in, icon figure is used directly for its better performance and preference compared with icons with background.