To build a harmonious and civilized Internet environment, poor text messages on the network to enhance the recognition and response capabilities. Article uses a novel method based on growing self-organizing feature map (GSOFM) and latent semantic indexing (LSI) method for performing a combination of text clustering.The combination of these two algorithms to find global and local features of the model. Experiments under the same conditions used in this new model and a single GSOFM and compared. Experimental results show that: The new combination of two technologies compared with the single GSOFM method improves the accuracy of clustering results, reducing the computation time for performing text clustering network provides a better way.%为建设和谐文明的网络环境,提升对网络不良文本信息的识别和应对能力.文章使用一种新颖的基于增长型自组织特征映射(GSOFM)和潜在语义索引(LSI)相结合方法用于不良文本聚类.这两种算法的结合能够发现全局和局部的模式特点.实验在相同的条件下使用了这种新颖的模式并和单一的GSOFM相比较.实验结果证明:这种新的两种技术的结合与单一的GSOFM方法相比提高了聚类结果的精确性,缩短了计算时问,为网络不良文本聚类提供了一种较好的方法.
展开▼
