E‐learning platforms are becoming more and more important and they are gradually changing people’s learning ways. In the e‐learning platforms, users actively create and join their favorite communities to share their questions and ideas. With the increase of users of e‐learning platforms, the number of communities is increasing dramatically. In this context, it has become difficult for users to find learning communities that match their interests and preferences. Therefore, how to effectively recommend the learning community for users has become an urgent need. However, compared to learning item recommendation, there is relatively limited work on learning community recommendation, and the existing research on community recommendation often ignores the personality information. Personality is considered one of the primary factors that influence human behavior and social relationships, as it affects how people react and interact with others. Several studies have demonstrated that people with similar personality tend to have similar interests. Furthermore, homophily theory also states that social interactions between similar individuals occur at a higher rate than among dissimilar ones. Since interests and interactions are important driving forces for users to join the learning communities, personality has an important impact on users’ choices of communities. Therefore, this paper aims at shedding some light on the impact of personality information on the accuracy of community recommendations. Particularly, we propose three enhanced matrix factorization models based on the Big Five personality framework. To evaluate the effectiveness of our proposed models, we conducted extensive experiments on myPersonality datasets. The results prove that the personality information can improve the performance of the learning community recommendation model and alleviate the data sparsity problem.

中文翻译：

---

利用个性信息来改善在线学习平台中的社区推荐

电子学习平台变得越来越重要，它们正在逐渐改变人们的学习方式。在电子学习平台中，用户积极创建并加入自己喜欢的社区，以分享他们的问题和想法。随着电子学习平台用户的增加，社区的数量正在急剧增加。在这种情况下，用户很难找到符合他们兴趣和爱好的学习社区。因此，如何有效地向用户推荐学习社区已成为当务之急。但是，与学习项目推荐相比，学习社区推荐的工作相对有限，现有的关于社区推荐的研究往往忽略了个性信息。人格被认为是影响人类行为和社会关系的主要因素之一，因为它影响人们的反应方式以及与他人的互动。多项研究表明，性格相似的人往往具有相似的兴趣。此外，同构理论还指出，相似个人之间的社交互动发生率高于异同人群之间。由于兴趣和互动是用户加入学习社区的重要驱动力，因此个性对用户对社区的选择产生重要影响。因此，本文旨在阐明人格信息对社区推荐准确性的影响。特别是，我们提出了基于“五种人格”框架的三种增强矩阵分解模型。为了评估我们提出的模型的有效性，我们在myPersonality数据集上进行了广泛的实验。结果表明，个性信息可以提高学习社区推荐模型的性能，减轻数据稀疏性问题。