基于加权频繁子图挖掘的图模型在文本分类中的应用

针对传统文本分类算法的分类精度低和计算复杂度高的问题,提出一种基于加权频繁子图挖掘的图模型文本分类算法。首先将文档集表示成图集;然后运用加权图挖掘算法提取频繁子图;最后,对特征向量进行分类。提出的算法仅提取最重要的子图,使其整体具有较好的分类效果和较高的计算效率。为评估该算法有效性,将其与多种现有分类算法分别对一个数据集进行分类实验,实验结果表明,提出的算法具有更高的识别精度和更少的运行时间。

Application of graph model based on weighted frequently subgraph mining in text classification

For the issue that the classification accuracy and the calculation of traditional paper classification algorithm complexity, put forward a text classification method based on graph model. This algorithm first document sets into the atlas, then the application of weighted graph mining algorithms to extract frequent subgraph, finally further processing to generate classification feature vectors for classification. Weighted subgraph mining algorithm to extract only the most important sub graph, the overall computational efficiency algorithm has better classification effect and higher. To evaluate this algorithm, the algorithm and the existing classification algorithms are classified in this paper experiments on a data set, the results indicate that the data sets the method has high accuracy, low operation time.