蚁群算法在文本聚类中的应用研究

Research of Text Clustering Based on Ant Colony Algorithm

摘要: 为提高文本聚类结果的精度和文本聚类的收敛速度, 对蚁群文本聚类算法进行了改进, 改进的措施主要包括修改迭代终止条件、动态调整蚂蚁观察半径、改变蚂蚁移动策略, 并且在复旦大学中文文本分类语料库上进行了仿真实验.实验结果表明, 改进后的蚁群文本聚类算法不仅加快了文本聚类的收敛速度, 而且提高了文本聚类结果的精度.

Abstract: In order to improve the accuracy of text clustering and clustering of convergence, Ant Colony clustering algorithm was improved. Measures for improvement consists of 3 main aspects, namely modified linear iteration stopping conditions, dynamic adjustment of ants observed radius, change ants moving strategy. An experiment was finished in Fudan University on the Chinese text corpus.Eexperimental results show that the improved Ant Colony clustering algorithm can speed up the convergence of the text clustering and improved the accuracy of text clustering.