摘要
知识发现是数字油藏的重要内容,也是建设数字油藏的主要目的之一。针对油气田开发的需要和油藏数据体的特点,本文综合利用数据清洗、数据挖掘、知识评估、知识解释、可视化等技术,提出了在数字油藏中进行知识发现的一种新思路,并用实例分析说明其实现方法,即以决策树技术分析油气田开发中采收率的影响因素为例,通过连续属性值的离散化处理和决策树的构建、剪枝以及知识评估和解释,从而达到准确、快速地挖掘出油藏数据库、油藏数据仓库和其它油藏数据体中大量有意义的规则、模式等知识。
Knowledge discovery is an important part of digital reservoir, and it is also one of the main objectives that can build digital reservoir. According to demand of oil and gas development and characters of reservoir data set, puts forward a new idea of knowledge discovery in digital reservoir by the technology, such as data cleaning, data mining, knowledge evaluation, knowledge interpretation, visualization, and illuminate the approach for its implication by the experiment, that is, analyzes factors influencing oil recovery efficiency by decision tree technology as example, through discretization of continuous attribute value, construction and pruning of decision tree and evaluation and interpretation of knowledge, and then discovers a great deal of interesting knowledge efficiently, such as rule and mode. from reservoir database, reservoir data warehouse and other reservoir data set.
出处
《微计算机信息》
北大核心
2007年第34期260-262,共3页
Control & Automation
基金
国家自然科学基金重大研究计划资助项目名称为基于数字岩心的孔隙网络建模方法研究(90610015)
关键词
数字油藏
知识发现
决策树
C4.5算法
digital reservoir, knowledge discovery, decision tree, C4.5 algorithm
