摘要
可视化挖掘是数据挖掘的重要研究领域,但目前的研究还主要集中在挖掘结果的可视化,用户对挖掘过程仍然缺乏可控性。本文利用focus+context技术实现探索型交互式数据挖掘。充分利用人类用户的认知能力与计算机的数据处理能力,实现用户对挖掘过程的参与从而利用用户领域知识于挖掘过程。适应人类的认知心理,便于帮助用户对挖掘结果的定位、聚焦、理解与评估,进而快速找到相对于当前应用上下文的有价值信息。并实现了一个原型系统IMARFC。
Abstract-Data mining is an iterative process to uncover novel interesting pattern from data. And human users play an important part for utilizing domain knowledge in the process. In this paper, a method is given for exploratory active mining of association rules based on a novel broad view and detail on demand style focus + context visualization technique. Specifically, the whole process is under the guidance of human users and such makes full use of human's cognitive capability and the computer's data processing power as well. A prototype system IMARFC is developed to show the feasibility our approach.
出处
《计算技术与自动化》
2007年第1期111-114,118,共5页
Computing Technology and Automation
基金
江西省自然科学基金项目(项目号:0411046)
江西省科技厅工业攻关项目(项目号:赣财教[2005]132号)