Decision-making layer model fusion method based on model expression clustering and short-distance estimation

The invention discloses a decision-making layer model fusion method based on model expression clustering and short-distance estimation. The method comprises the following steps: A, obtaining a training set; B, obtaining an identification model; C, obtaining model parameter expression; D, obtaining a clustering sample; E, obtaining model parameter expression of the to-be-predicted sample; F, calculating the proximity of the to-be-predicted sample to the clustering sample in the training set in the recognition model; G, calculating the credibility of each identification model; H, calculating fusion training sample expression; I, calculating fusion test sample expression; J, training a fusion model on the fusion training sample expression by using a classification model algorithm; K, carryingout classification and identification on fusion test sample expression by utilizing the obtained fusion model. The fusion effect can be effectively improved. 本发明公开了一种基于模型表达聚类和近距性估计的决策层模型融合方法，包括以下步骤：A：获得训练集；B：获得识别模型；C：获得模型参数表