一种有效的基于non-metric距离的聚类模型
An Effective Clustering Model Based on Non-metric Distance
-
摘要: 详细分析了non-metric距离对传统聚类算法的影响，提出了用有向图和有向树描述的聚类模型．基于该模型，给出了能对具有non-metric距离特征的数据进行有效聚类的算法．在实际的购物篮数据集上进行了验证；实验结果表明，所提算法在保证聚类效率的同时，大幅度提高了聚类质量．Abstract: The influence of non-metric distance on classical clustering algorithms is discussed in detail,and a clustering model based on directed-graph and directed-tree is proposed.Based on the model,an effective clustering algorithm for the datasets with non-metric distance measure is presented.Verifications are made on a real basket dataset,and the experimental results show that the proposed algorithm improves the clustering quality and efficiency.
下载:
