| 联系发现：一种新的数据挖掘方法综述 |
| |
| 引用本文： | 陈飞,商琳,骆斌,陈世福.联系发现：一种新的数据挖掘方法综述[J].计算机科学,2006,33(11):123-127. |
| |
| 作者姓名： | 陈飞 商琳 骆斌 陈世福 |
| |
| 作者单位： | 南京大学计算机软件新技术国家重点实验室,南京,210093 |
| |
| 摘 要： | 联系发现是数据挖掘中较新的研究领域。联系发现是一种对海量数据进行挖掘，找出其中潜在模式，抽取有用知识并发现隐藏联系的技术。本文首先综述了联系发现的概念、范围、特点和难点等，详细介绍了联系发现的几种主要方法：无监督的联系发现方法（新颖联系发现）、使用归纳逻辑程序技术挖掘关联数据的联系发现方法、多假设反演推理的联系发现方法、基于相关分析的联系发现方法以及KOJAK组队探测器，讨论了联系发现系统性能评估的方法与联系发现的置信区间度量方法，并简要描述了联系发现的一个具体应用的实例一——证据抽取和联系发现研究计划（EELD），最后探讨了目前联系发现研究中出现的问题及未来发展趋势。
|
| 关 键 词： | 联系发现 数据挖掘 反演推理 相关分析 性能评估 |
Link Discovery: A New Data Mining Method |
| |
| CHEN Fei,SHANG Lin,LUO Bin,CHEN Shi-Fu.Link Discovery: A New Data Mining Method[J].Computer Science,2006,33(11):123-127. |
| |
| Authors: | CHEN Fei SHANG Lin LUO Bin CHEN Shi-Fu |
| |
| Affiliation: | National Laboratory for Novel Software Technology, Nanjing University, Nanjing 210093 |
| |
| Abstract: | Link Discovery is a relatively new form of data mining. Link Discovery is the technique for mining large amounts of data to find hidden patterns, extract valuable knowledge and discover hidden links. A survey on link discovery is presented in the paper, including the definition, the research range, characteristics and challenges of Link Discovery. And several typical Link Discovery methods are detailed, including unsupervised Link Discovery methods (novel Link Discovery), relational data mining with inductive logic programming for Link Discovery, multi-hypothesis abductive reasoning for Link Discovery, Link Discovery based on Correlation Analysis and the KOJAK group finder. We emphasize on the performance evaluation metrics for Link Discovery systems and measuring confidence intervals in Link Discovery. Further more the aplication of Link Discovery is presented: EELD. Lastly some issues about Link Discovery are discussed and the future directions of Link Discovery are pointed out. |
| |
| Keywords: | Bayesian learning Reinforcement learning Single-agent Multi-agent |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机科学》浏览原始摘要信息 |
|
点击此处可从《计算机科学》下载全文 |
|
