Data-driven modeling and predicting method for buoy motion characteristics

The invention provides a buoy motion characteristic data-driven modeling and prediction method, which comprises the following steps of: (1) determining input and output variables of a prediction model, and collecting a modeling sample set S = {X, Y}; (2) establishing a local LSSVR model of buoy motion characteristics: (2.1) automatically dividing a sample set S into two sample subsets, namely S1 under a normal working condition and S2 under an extreme working condition; (2.2) independently carrying out learning training on each sample subset, and establishing local LSSVR prediction models LSSVR1 and LSSVR2 of the buoy motion characteristics; and (3) automatically selecting a proper local LSSVR prediction model for each new test sample. The beneficial effects of the invention are that the data-driven modeling and prediction method provided by the invention can achieve the modeling and prediction of the motion characteristics of the buoy in the complex frequency-varying marine environment based on a limited mod