Tree similarity calculation method based on structural features

In order to effectively calculating the similarity between trees, the invention provides a method for calculating the similarity between trees through structural features of a tree T. All subtrees in the non-isomorphic form, which are provided with K nodes, are configured, the number of the isomorphic subtrees in the subtrees is calculated from the T, and the number is taken as a feature vector to carry out similarity calculation of the tree. The algorithm has important research significance on the structural feature extraction, similarity calculation and clustering of a tree-like structure and even the graph mining field. 为了高效计算树的相似度，本发明提出通过树T的结构特征计算树的相似度的方法，利用构造K个节点的所有非同构形态的子树，从T中计算这些子树的同构个数，将其作为特征向量来进行树的相似度计算。本算法对于树形结构乃至于图挖掘领域的结构特征提取、相似度计算和聚类都有重要的研究意义。