Knowledge Discovery Technology Research Of Exploratory Well Production Management Decision Support System

Exploratory Well Production Management Decision Support System tried to apply artificial intelligence techniques to achieve computer intelligence decision, which is used to improve the veracity and speed of decision of exploratory well production management and reduce cost of exploratory well production. In the process of system developing, knowledge discovery techniques are applied to gain useful knowledge and rule from large of exploratory data, as the supplementary of incomplete knowledge and experience of experts.This paper analyzes data characteristic and feature based on Exploratory Well Production Management Decision Support System, and chooses a more appropriate data pretreatment method. The low accurate calculation rate, multi-value orientation and other questions occur when existing decision tree algorithms apply to system. Combining with rough set theory, this paper proposes two decision tree algorithms. One bases on attribute weigh and the other bases on rough classification degree. Two algorithms change existing decision tree algorithms from different perspectives. Decision tree algorithm based on attribute weigh makes use of attribute reduction to weigh, according to different impact on decision of different attributes. In the process of constructing decision tree, the attribute which has the largest weight is a splitting property. Decision tree algorithm based on degree of rough classification considers the impact on decision results of attribute classification accuracy and dependency between attributes.The experiment proves that two decision tree algorithms effectively improve the classification accuracy and decision veracity.