项目名称: 基于分形与数据流挖掘技术的动态数据挖掘方法及其应用研究
项目编号: No.61202227
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 计算机科学学科
项目作者: 刘慧婷
作者单位: 安徽大学
项目金额: 24万元
中文摘要: 该项目主要研究在异质异构商务数据资源(包括客户信息、在线交易数据、日志数据流)中,利用分形和数据流挖掘相关技术进行动态数据分析的基本理论与方法。 项目利用EMD方法滤除商务数据的噪声,实现商务数据的预处理,为后续的商务数据分析工作提供干净的数据,提高分析结果的可信度。 同时,项目使用分形技术分析复杂商务数据中数据的分形特征,以此研究在动态环境下分形知识的发现过程;利用数据流挖掘技术动态挖掘单商务流中的潜在规则和多商务流之间的相似性,以进一步用于顾客个性化推荐服务、商家推销和进货策略的改进。 该研究可望解决商务数据分析中噪声干扰和如何实现动态分析的问题,提高商务智能系统的智能程度和复杂环境下的时效性,从而能更好地为商务决策提供及时的支持。
中文关键词: 不确定数据;数据流;频繁项集;个性化推荐;
英文摘要: The project mainly researches on the basic theories and methods of dynamic data analysis using fractal and data stream mining technology, and the data is business data in heterogeneous resources (including customer information, transaction data, log data flow). The project makes use of the EMD method to filter out noise of business data, and this can realize business data pre-processing, provide clean data for the follow-up analysis of business data, increase credibility of results of the analysis. At the same time, the project uses fractal technology to get fractal characteristics of complex business data, and this can study fractal knowledge discovery process in a dynamic environment. Use data stream mining technology to dynamically mine of potential rules a single business flow and similar nature of multi-business flow, and these can be further utilized for customer personalized recommendation service, improvements of the policy of businesses sell and stock. The study is expected to solve the noise problem in the analysis of business data and know how to implement dynamic analysis, improve the degree of intelligence and efficiency of business intelligent system in complex environment, and thus better able to provide timely support for business decision-making.
英文关键词: Uncertain data;data stream;frequent itemsets;personalized recommendation;