论文部分内容阅读
为了探讨新的加权系数估计方法对于消除或减弱证据层不满足条件独立性假设时对预测结果的影响,对加权证据权模型的加权系数估计方法进行了新的探讨,尝试用顺序估计法估计加权系数.加权系数的顺序估计法是将加权证据权模型与基于模糊预测对象的证据权模型相结合,将证据层按照一定顺序逐步加入到加权证据权模型中,在加入到模型的过程中依次用已经获得的后验概率作为模糊训练层对证据层加入到模型的顺序进行修正,并通过条件相关系数的方法估计加权系数.分别以1组多元正态分布模拟数据和个旧锡铜多金属矿产资源预测为例,比较了多种模型的后验概率,结果表明加权证据权模型对减弱证据层不满足条件独立性假设所产生的影响是有效的.
In order to explore how the new weighted coefficient estimation method can eliminate or weaken the influence of the evidence layer on the prediction results when the conditional independence assumption is not satisfied, a new approach to estimate the weighted coefficient of the weighted evidence weight model is proposed. Coefficient.The sequence estimation method of weighted coefficient is to combine the weighting evidence weight model and the evidence weighting model based on the fuzzy prediction object and gradually adding the evidence layer to the weighting evidence weighting model in a certain order, The acquired posterior probability is used as the fuzzy training layer to modify the order of the evidence layer added to the model and the weighted coefficient is estimated by the method of conditional correlation coefficient.According to 1 multivariate normal distribution simulation data and Gejiu tin-copper polymetallic mineral resources For example, the posterior probabilities of multiple models are compared. The results show that the weighted weight model is effective in reducing the impact of the evidence layer not satisfying the conditional independence assumption.