Mining Method of Microbial Community Characteristics in Hangzhou Xixi Wetland

Aiming at the problems of low mining accuracy and long mining time in traditional wetland microbial community feature mining methods, a microbial community feature mining method in Hangzhou Xixi Wetland is proposed. The improved local linear embedding (MLLE) method was used to reduce the dimensionality of the sample data of the microbial community in Xixi Wetland and to extract the maximum weight of microbial community characteristic types in Xixi Wetland. The association rule evaluation method was used to determine the abnormality of the microbial community feature vector and remove the redundancy in the feature sequence. On the basis of removal processing, based on the feature mining strategy, the feature subset was mined, and the pros and cons are evaluated to obtain the optimal microbial community feature subset. The simulation results show that the proposed method for mining microbial community features in Xixi Wetland in Hangzhou has higher precision and faster mining efficiency.