摘要
应用基于粗糙集理论的数据挖掘技术对数量巨大的钻头数据资料进行处理,能在保留关键信息的前提下,对钻头数据进行约简并求得知识的最小表达,去除冗余信息。这样在使用人工神经网络优选钻头时,网络的训练样本数和训练步数都有较大减少,而迭代精度却明显提高;实例计算表明,使用约简后的样本数据进行钻头选型,优选的钻头序列更加合理;由于钻井行业涉及的数据量巨大,这种数据挖掘技术应该得到足够重视。
The huge bit data are processed with the data mining based on the rough sets theories. This method can obtain minimal expression of the bit datum with maintaining key information. The nerve network trining data will decrease largerly and the iterating accuracy degree of nerve network will increase greatly at the same time. The results of caculating show that the seleting bits with the data after the processing are more reasonably than prvious bits. The data mining method should be counted because the data are huge in the field of drilling.
出处
《断块油气田》
CAS
2007年第6期60-62,共3页
Fault-Block Oil & Gas Field