The embodiment of the invention provides a soil bacterial
community structure evaluation method and
system based on an intermediate
infrared spectrum
transfer function and application, and relates to the technical field of soil
microbial ecology. The method comprises the following steps: S1, collecting a soil sample; s2, acquiring a mid-
infrared diffuse reflection spectrum of the soil sample to be detected; s3, acquiring an environment covariant corresponding to the soil sample, wherein the environment covariant comprises at least one of soil physicochemical properties, climate data and topographic parameters; s4, splicing the mid-
infrared diffuse reflection spectrum obtained in the step S2 and the environment covariable obtained in the step S3 to form a fusion
feature vector as an input
feature set; and S5, inputting the input
feature set into a pre-trained nonlinear
machine learning regression model, and outputting the relative abundance and / or OTU richness of at least one
bacteriophage in the soil sample. According to the method, a technical
route of combining a mid-infrared (MIR) spectrum and a spectral
transfer function (STF) is adopted and is used for predicting the soil bacterial
community structure.