【摘要】 在油、气预测中,并非用到的地震属性越多,预测效果就越好。由于计算误差等影响,冗余地震属性反而会使预测的准确率降低。目前地震属性约简较多采用粗糙集的思想优化地震属性。然而,等价关系在有些情况下难以实现,而且连续地震属性经离散化处理后会造成原始数据失真,因此本文提出基于覆盖粗糙集的地震属性约简方法,不仅有效解决了地震属性等价关系和数据失真的问题,还克服了粗糙集理论在地震属性约简中应用的局限性,使粗糙集理论更具一般化。仿真试验和实际应用表明,通过样本地震属性约简,可以提高油、气预测精度。更多还原

【Abstract】 In oil & gas prediction,it is not always true that the more the index variables of seismic attribute data are,the better effect of prediction is.On the contrary,the classification accuracy will reduce by the redundant index variables because of the calculation error.Nowadays,rough set is mostly applied to optimize seismic attributes in the seismic attribute reduction.However,it is difficult to achieve the equivalence relation in some cases,and the process of discretizing continuous attributes can lead to the distortion of the original data.Therefore,the seismic attribute reduction based on the covering rough set is presented in this paper.This approach not only solves problems mentioned above,but also overcomes limits in rough set application,and makes the rough set theory more generalized.The simulation and actual experiments reveal that the results of the seismic attribute reduction application can improve oil & gas prediction precision.更多还原