人工神经网络自组织分类解释模型研究

应用人工神经网络技术，实现数据的自组织分类；针对不同的类，建立了相应的解释模型。实际计算时，采用统计模式识别方法选择解释模型。以渗透率为例，比较了经验公式、人工神经网络方法和人工神经网络自组织分类方法的计算结果，表明由经验模型到神经网络单一混合模型再到神经网络组织分类模型，其平均相关系数逐步提高，平均绝对误差、平均相对误差依次降低。证明自组织分类方法具有更大的优越性。

Establishing Interpretation Models of Self-organizatiom Classification

Neural network technique is used in self organization classification of data and appropriate interpretation models are established for various classes In actual calculation, interpretation models are selected by using the result of statistic pattern recognition By comparing the result from empirical formula, neural network method and self organization classification, it is found that the self organization classification in logging interpretation is superior to other methods in that:the average correlation coefficient is increasing and the average absolute error and relative errors are decreasing, correspondingly