Document Classification by Semi-supervised Online Learning Based on ART
-
摘要: 根据自适应谐振理论提出了半监督学习自适应谐振理论系统.在该系统中取消了一般半监督学习算法中假定已知数据概率分布的条件限制,利用自适应谐振理论的稳定性和可塑性,使其具有非常强的学习新模式和纠正错误能力.为了提高系统自适应性能力,将警戒参数设置为动态变化。实验结果表明半监督学习自适应谐振理论系统的性能优于判别式CEM算法,特别是在含有噪音和新模式数据情况下,其优势更为明显.Abstract: A semi-supervised learning system was proposed based on ART(adaptive resonance theory).It overcomes the limitation in the assumption in other semi-supervised learning algorithms that probabilistic distribution of data is known,and has the strong ability of learning new patterns and correcting errors because of stability and plasticity of the adaptive resonance theory.Higher adaptability of the system was advanced by setting vigilance parameters dynamically.Experimental results illustrate that the performances of the proposed system is better than the discriminant CEM(classification expectation maximization) algorithm,particularly when there are noise data and new patterns.
-
Key words:
- on-line learning /
- document classification /
- ART /
- semi-supervised learning /
- vigilance parameter
-
ZHANG T,OLES F.A probability analysis on the value of unlabeled data for classification problems[C] ∥Proc.Intl Conf.Machine Learning (ICML).San Francisco:Morgan Kaufmann,2000:1 191-1 198.[2] COHEN I,COZMAN F G,BRONSTEIN A.On the value of unlabeled data in semi-supervised learning based on maximum-likelihood estimation[R]. Technical Report HPL-2002-140,Hewlett-Packard Labs,2002.[3] CARPENTER G A,GROSSBERG S,REYNOLDS J H.ARTMAP:supervised real-time learning and classification of non-stationary data by a self-organizing neural network[J]. Neural Networks,1991,4:565-588.[4] CARPENTER G A,GROSSBERG S.A massively parallel architecture for a self-organizing neural pattern recognition machine[J]. Computer Vision,Graphics,and Image Processing,1987,37:54-115.[5] KOLLER D,SAHAMI M.Hierarchically classifying documents using very few words[C] ∥Proc.ICML-97,Nashville:Morgan Kaufmann,1997:170-176.[6] NOEL V J,REZA A M,PATRICK Gallinari.Learning classification with both labeled and unlabeled data[C] ∥Proc.European Conference on Machine Learning,Lecture Notes in AI.Helsinki:Springer,2002:468-479. -
点击查看大图
计量
- 文章访问数: 1452
- HTML全文浏览量: 64
- PDF下载量: 203
- 被引次数: 0
下载: