基于Bayes逐步判别分析的油气储量价值分级

油气储量是国家的战略资源，鉴于其经济性差异，要实现资产化管理，必须首先评价其价值的优劣等级。本文选取影响油气储量价值等级的7个因素，将Bayes逐步判别分析方法应用于油气储量价值等级评价与分类中，建立了油气储量价值综合评判的Bayes判别分析模型。首先，经逐步分析，剔除储层渗透率，模型筛选出采收率、储量丰度、原油粘度、储层埋深、储量规模、原油凝固点等6个指标作为有判别意义的判别因子；然后，以已知价值等级的油气储量数据作为训练样本。将油气储量价值划分为优、良、中、差4个等级，建立4个Bayes线性判别函数，并回判检验判别函数的优良性。最后，并对待判样品进行Bayes判别。研究表明，Bayes逐步判别分析模型误判率较低，回判正确率达到85％，是评价油气储量价值等级的有效方法。

Application of Bayes Stepwise Discriminatory Analysis in Value Classification of Oil&gas Reserves

Oil＆gas reserves are the strategic resources. In order to carry our capitalization management, value classification of oil＆gas reserves has to be assessed due to economical deference. Seven factors influencing value classification of oil＆gas reserves are chosen. Bayes stepwise discriminatory analysis is applied to classify reserves value of oil＆gas. Based on it, Bayes analysis model of value classification of oil＆gas reserves is established. Firstly, through stepwise analysis, six indicators are chose as significant discriminatory factors, which are recovery ratio, reservoir abundance, crude oil viscosity, reservoir depth, reservoir scale and crude oil freezing point. And one factor, that is permeability, is eliminated. Secondly, based on developed reservoir data of oil＆gas, reservoir value can be classified into four gradations： excellent, good, medium and bad, which are regarded as four normal populations of Bayes discrimination. Through multivariate statistics, four Bayes linear functions are gained. Finally, pending test samples of oil＆gas reserves are discriminated by the functions. It is found that the mistake rate of Bayes stepwise model is low, which demonstrates that Bayes stepwise discriminatory analysis is feasible in value classification of oil＆gas reserves.