基于蚁群算法的文本分类和聚类

为了研究并提高文本的分类和聚类算法的性能，笔者根据蚁群算法在TSP问题中的应用方法，将其改进引用到文本的分聚类中。在文本聚类中，改变蚂蚁的信息素释放机制，道路节点的聚合方式，最终将相似文本进行聚合。在文本的分类中，将所需要的分类信息装入蚂蚁，蚂蚁根据系统外部所希望的方式将文本分类。实验结果证明，这种新的算法可以使文本分类和聚类的准确度提高，蚁群算法在文本分类聚类中的应用是可行的。

Text Classification and Clustering Based on Ant Colony Algorithms

In order to study and improve performance of text classification and clustering, the authors, based on the usage of ant colony algorithm in solving the TSP（travelling salesman problem）, modify and use this algorithm in the text classification and clustering. When this algorithm is used to cluster texts, the way for releasing ants＇ pheromone, and the mode for clustering path-nodes as well should be changed, and finally the similar texts are placed together. In text classification, the information must be told to the ants, which indicates the final categories and is wanted before the process. The experiment indicates the facts that this new method could increase the rate of accuracy, and that the ant colony algorithm could be used in text classification and clustering.