论文部分内容阅读
刘堂宴 ,郑勇 ,傅容珊 ,王绍民 .提高测井建模精度的一种方法 .石油地球物理勘探 ,2 0 0 2 ,37(1) :44～ 47本文提出了测井建模时利用反演理论的 L1范数对个别离群点的不敏感特性删除离群点数据的一种新方法。该方法首先用 L1范数建立方程参数的初始值 ,再用最大距离迭代方法去掉离群点 ,最后利用高斯假设 (最小二乘法 )在没有离群点的数据中建立测井模型。应用效果表明 ,该方法可实现离群点数据的自动删除 ,并具有客观删除标准。应用这一方法建立的测井评价模型满足精度要求 ,操作规范
Liu Tang-fei, Zheng Yong, Fu Rong-shan, Wang Shao-min.A Method to Improve the Precision of Well Logging.Petroleum Geophysical Prospecting, 2002, 37 (1): 44 ~ 47 In this paper, A New Method of Deleting Outlier Data for Insensitive Features of L1 Norms for Individual Outliers. Firstly, the initial value of the equation parameter is established by using L1 norm, then the outlier is removed by the maximum distance iteration method. Finally, the log model is established by using Gaussian assumption (least squares method) in the data without outliers. The application results show that this method can automatically delete the outlier data and has the objective deletion criterion. The logging evaluation model established by this method meets the precision requirements and operation specifications