| 摘 要: 目前小说的受众群体越来越大，其中蕴含着巨大的商业价值。文本聚类的研究领域也在突飞猛进，但对 于其中的现实领域：小说聚类，相关的研究却较少。本文研究了一种基于小说中的社交网络对其进行聚类的方法。该方 法首先提取出小说中的社交网络，在得到网络的特征向量后，基于其进行聚类，并将结果与依据小说作者的划分进行 对比。实验结果表明，该方法可以在一定程度上反映出不同作者写作风格的不同，效果可以接受，并拥有进一步提升 的可能。 |
| 关键词: 小说；社交网络；聚类算法 |
|
中图分类号: TP391.1
文献标识码: A
|
|
| Clustering of Novels Based on Social Network |
|
LOU Kaiyi,BA Yuanjie,LI Shaoang
|
( School of Computer Science, Jilin University, Changchun 130012, China)
|
| Abstract: At present,more and more people are reading novels,which contains great commercial value.The research field of text clustering is also advancing by leaps and bounds,but for the real practice—novel clustering,there are few related researches.This paper uses a method based on social network in the novel to cluster it.The method first extracts the social network in the novel.After obtaining the feature vector of the network,it clusters based on it and compares the result with the division according to the author of the novel.The experimental result shows that the method can reflect the different writing styles of different authors to a certain extent,the effect is acceptable,and further improvement is possible. |
| Keywords: novels;social network;clustering algorithm |
