摘要
针对传统模糊认知图和知识地图数据挖掘效率偏低且预测准确性不高的问题,提出了基于智能知识地图的数据挖掘方法。利用成对变量的离散、映射从原始数据构建散布图;定义四元组模糊关联规则,在此基础上构建智能知识地图;根据关联节点数目识别中枢节点和不活动节点。实验利用静态分析评估了系统结构的不确定性,通过时域分析揭示了知识地图主要属性的演变过程,对上证50指数公司的金融数据分析验证了所提方法的有效性,实验结果表明,所提方法在金融危机预警方面取得非常准确的预测效果,为决策者掌控公司运营状况提供了强有力的危机预警工具。
For the issue that the traditional fuzzy cognitive map and knowledge map have inefficient data mining and low pre dictability, data mining based on intelligent Knowledge Map (KM) is proposed. Paired variables are used to construct scatter dia gram by discrete and mapping from original data. Quad fuzzy association rules are defined based on which knowledge map is constructed. Numbers of linked nodes are used to recognize central node and inactive nodes. Uncertainty of the system structure is estimated by static analysis, and evaluation process of major KM attributes is uncovered by timedomain analysis. Then effi ciency of proposed method has been verified by analysis on financial data of 50 index companies in Shanghai Stock. Experi ments results show that the proposed method has perfect predicting effect on financial crisis warning, which provides a powerful crisis warning tool for deciders controlling operating condition of company.
出处
《计算机工程与应用》
CSCD
2013年第24期116-121,共6页
Computer Engineering and Applications
基金
福建省科技支撑计划项目(No.102102210419)
关键词
智能知识地图
数据挖掘
金融危机预警
静态分析
时域分析
intelligent knowledge map
data mining
financial crisis warning
static analysis
time-domain analysis