测井数据模式识别中的信息优化方法

简介了智能信息处理中新出现的ＲｏｕｇｈＳｅｔ（ＲＳ）理论，探讨了测井数据模式识别中的信息优化原理，提出了将ＲＳ理论与人工神经网络结合起来进行测井信息优化与模式分类的ＲＳ神经网络智能系统，并将该系统应用于测井样品分类，分类结果表明，本系统速度快，易实现，而且在优选测井属性，最大程度地减少使用测井属性种数，提高分类正确率等方面，明显优于其它方法，本系统为测井属性选择提供了一条新途径，将成为测井数据模式

A Method for Information Optimization in Pattern Recognition With Log Data

First, briefs the Rough Set (RS) theory, which is a new theory in intelligent data processing; then disscusses the principle of information optimization in pattern recognition with log data; and finally advances an intelligent system combining RS with artificial neural network, which is used for log information optimization, pattern recognition and log samples classification. The result of classification shows that the system has the advantage of fast computation and easy realization, and is better than other methods in optimizing log attributes, decreasing the indexes of log data to a great extent, and enhancing the correct ratio of classification, etc. Therefore, the system provides a new way of log attribute selection and will become an effective means of pattern recognition with log data.