Aiming at the roughness of theme category existing in question and answering(Q&A) community, using particle swarm optimization algorithm, concepts of community seeds and community themes are introduced. Firstly, explicit linkages existing in Q&A community are mined and the basic structure of Q&A community is built. Then, the contents of Q&A community are deeply analyzed, and the topics of Q&A community are defined according to the recessive characteristics of question nodes. The theme categories are refined until the structure is stable. Experimental results show that the algorithm can accelerate the question nodes’ convergence and greatly improves the accuracy of theme mining of Q&A community.% 针对问答社区中问题主题类别划分的粗糙性,应用粒子群优化算法,引入问答社区种子和问答社区主题的概念,首先挖掘问答社区中存在的显性联系,构建基本问答社区结构,然后,深入分析问答社区内容,根据问题节点之间的隐性特征,定义问答社区主题,精分细化问答社区主题类别,直到结构稳定。实验结果表明,该算法能加速问题节点的收敛,极大地提高了问答社区主题挖掘精度。
展开▼