中文摘要：

随着社会媒体的普及,用户信息的爆炸式增长为深入理解在线用户行为提供了非常丰富的信息源.由于用户人格特质是用户行为的主要驱动力,人格特质的差异可能会对用户的在线行为产生一定的影响,因此,用户人格特质识别问题近年来受到了众多学者的关注.首先,基于用户网络结构信息和用户发布内容信息序列构建用户人格特质识别特征,并根据特征重要性为其分配权重.然后,以用户人格特质相关因子约束目标函数,从用户社会网络结构特征、语言学特征和情感特征三个维度利用非负矩阵分解方法识别社会网络中用户的五大人格特质.最后,在真实的数据集上验证了提出框架的有效性,并通过实验以更细的粒度进一步验证了用户人格特质之间相关性的存在,同时证明了特征权重和用户人格特质间的相关性在用户人格特质识别问题中的重要性.文中为社会网络中的多维用户人格特质识别问题提供了一种新思路.

英文摘要：

With the pervasiveness of social media, the explosion of users＇ generated data provides a potentially very rich source of information for online researchers understanding user＇s behaviors deeply. Since user＇s personality traits are the driving force of user＇s behaviors and individual differences in user＇s personality traits may have an impact on user＇s online activities, as a consequence, user＇s personality traits recognition has attracted increasing attention in recent years. On the basis of user＇s network structure information and series of posts information, we first build user＇s personality traits recognition features, followed by distributing weights to features according to their different importance. And then, we utilize nonnegative matrix factorization to recognize user＇s Big Five personality traits from his/her network structure features dimension, linguistics features dimension and emotion features dimension by employing personality traits correlation factor to constrain objective function. Experiments on real-world Faeebook dataset demonstrate the effectiveness of the proposed framework. Further experiments are conducted not only to validate the existence of the correlations between user＇s personality traits from a more fine-grained view, but also understand the importance of different feature＇s weight and the importance of the correlations between user＇s personality traits in recognizing user＇s personality traits. What＇s more, we provide a new train of thought for multidimensional personality traits recognition in social networks.