随机森林回归在地震储层预测中的应用

发布时间：2018-01-21 21:40

本文关键词： 随机森林回归地震属性特征参数储层预测　出处：《石油地球物理勘探》2016年06期 　论文类型：期刊论文

【摘要】：针对储层预测的复杂非线性及稳定性问题,将随机森林回归算法引入到地震储层预测中,建立地震属性与储层特征参数之间的非线性关系。以多种不同的地震属性为基础,通过构建井旁道地震属性与特征参数的回归森林模型进行储层预测,以预测值与实际值之间的均方根误差值为评价标准,分析随机森林回归算法在地震储层预测中的特点。将本方法应用于某陆地工区的自然电位预测和某海上工区的自然伽马预测,并与支持向量回归机方法的预测结果进对比,结果表明,尽管地震数据受到较强噪声的影响,随机森林方法仍可较好地刻画出储层的三角洲前缘沉积特征,表现出较好的稳定性和较高准确性。
[Abstract]:Aiming at the complex nonlinear and stability problems of reservoir prediction, stochastic forest regression algorithm is introduced into seismic reservoir prediction. The nonlinear relationship between seismic attributes and reservoir characteristic parameters is established. Based on different seismic attributes, reservoir prediction is carried out by constructing a regression forest model between seismic attributes and characteristic parameters. The root mean square error between the predicted value and the actual value is taken as the evaluation criterion. The characteristics of stochastic forest regression algorithm in seismic reservoir prediction are analyzed. The method is applied to the prediction of natural potential in a land area and natural gamma prediction in an offshore area. Compared with the prediction results of support vector regression machine, the results show that, although the seismic data are affected by strong noise, the stochastic forest method can well depict the delta front sedimentary characteristics of reservoir. Show good stability and high accuracy.
【作者单位】：中国石油大学(华东)地球科学与技术学院;海洋国家实验室海洋矿产资源评价与探测技术功能实验室;中国石油大学(华东)信息与控制工程学院;
【基金】：国家自然科学基金项目(41674125)资助
【分类号】：P631.44
【正文快照】： 1　引言为了克服单一模型的局限性,集成的思想被逐渐被引入到数据挖掘技术中[1],人们提出了一系列的算法和方法,如Boosting方法[2]、AdaBoosting算法[3]、Bagging方法[4]、随机决策森林法[5]、随机子空间方法[6]、随机分割选取法[7]等。随机森林法(Random　Forests,RF)于2001

【相似文献】