| 摘要: |
| 使用集成学习技术来提高聚类性能.由于聚类使用的训练样本缺乏期望输出,与监督学习下的集成相比,在对个体学习器进行结合时更加困难.通过对不同的聚类结果进行配准,并基于互信息权进行个体学习器的选择,提出了基于Bagging的选择性聚类集成算法.实验表明,该算法能够有效地改善聚类结果. |
| 关键词: 机器学习 集成学习 聚类 非监督学习 选择性集成 |
| DOI： |
| 分类号: |
| 基金项目:Supported by the National Outstanding Youth Foundation of China under Grant No.60325207(国家杰出青年科学基金) |
|
| Bagging-Based Selective Clusterer Ensemble |
|
TANG Wei,ZHOU Zhi-Hua
|
| Abstract: |
| This paper uses ensemble learning technique to improve clustering performance. Since the training data used in clustering lacks the expected output, the combination of component learner is more difficult than that under supervised learning. Through aligning different clustering results and selecting component learners with the help of mutual information weight, this paper proposes a Bagging-based selective clusterer ensemble algorithm. Experiments show that this algorithm could effectively improve the clustering results. |
| Key words: machine learning ensemble learning clustering unsupervised learning selective ensemble |
