构造性覆盖算法的知识发现方法研究
Study on the Method of Knowledge Discover Based on the Structured Covering Algorithm
-
摘要: 该文提出一种新的基于构造性覆盖算法的知识发现方法。由于覆盖网络构造方法的特殊性,使得形成的每个覆盖领域都很有价值,对覆盖领域内样本分析能挖掘出数据内在的知识,且可以根据需求构造不同的覆盖网络,形成对数据的多侧面的分析;实验结果表明覆盖算法用于知识发现的方法是有效可行的。Abstract: This paper proposes a new method of knowledge discovery based on the structured covering algorithm. Since the network of covering domains is constructed through a special method, it makes each covering domain a valuable pattern. Through analyzing the samples covered by the covering domain, certain valuable pattern will be found, which includes clustering information of samples, association rules among the data, the outlier analysis, etc. And in order to meet different requirement of users, different covering networks can be structured, therefore, multi-sides of data can be analysed. The experiments show that using the covering algorithm to discover knowledge of data is effective and feasible.
-
Zhang Ling, Zhang Bo. A geometrical representation of McCulloch-Pitts neural model and its applications[J].IEEE Trans. on Neural Networks.1999, 10(4):925-929[2]张铃, 张钹, 殷海风. 多层前向网络的交叉覆盖算法. 软件学报, 1999, 10(7): 737-742.[3]吴鸣锐. 大规模模式识别问题的分类器设计研究. [博士论文], 北京, 清华大学计算机系, 2000.[4]陶品, 张钹等. 构造型神经网络双交叉覆盖增量学习算法. 软件学报, 2003, 14(2): 194-201.[5]叶少珍, 张钹等. 一种基于神经网络覆盖构造算法的模糊分类器. 软件学报, 2003, 14(3): 429-434.[6]McCulloch W S, Pitts W. A logical calculus of the ideas immanent in nervous activity[J].Bulletin of Mathematical Biophysics.1943, 5:115-133[7]张燕平, 张铃等. 基于覆盖的构造学习算法SLA及其在股票预测中的应用. 计算机研究与发展,2004, 41(6):979-984.[8]张燕平, 张铃, 吴涛. 机器学习中的多侧面递进算法MIDA. 电子学报, 2005, 33(2): 327-331.
计量
- 文章访问数: 2225
- HTML全文浏览量: 92
- PDF下载量: 714
- 被引次数: 0