欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
基于粗集和决策树两种方法的各自优势互补,提出将粗集与决策树相结合的新方法,并将此算法运用到个人信用卡发放模型中。首先利用布尔推理算法将连续属性进行离散化处理,然后采用一种以加权和属性重要度为启发信息进行属性约简,得到降维数据,最后采用J48决策树算法,得到决策规则。通过对比K最近邻分类、朴素贝叶斯、RBF神经网络、支持向量机等算法,这种新的数据挖掘算法保留了原有数据特点,加快了知识获取的进程,提高了模型的交叉验证率,简化了规则,取得了满意的研究结果。
英文摘要：
Based on the complementary advantages of rough set and decision tree method,put forward a new method of combining rough set and decision tree,and use this algorithm to personal credit card payment model. Firstly,use Boolean reasoning algorithm to continuous attribute for discretization processing,and apply a heuristic information of weighted and attribute importance for attribute reduction to obtain data with dimensionality reduction,finally utilize J48 decision tree algorithm to get the decision rules. Compared with the K nearest neighbor classification,Naive Bayes,RBF neural networks,support vector machines and other types of algorithms,this new data mining algorithm retains the original data characteristics to accelerate the process of knowledge acquisition,improving cross-model verification rate,simplifying the rules,getting the satisfactory results.
同期刊论文项目
同项目期刊论文
Several teleportation schemes of an arbitrary unknown multi-particle state via different quantum cha
Quantum three-qubit w-state sharing via positive operator-valued measurement and projective measurem
Remote information concentration via four-particle cluster state and by positive operator-value meas
位置：