A-KNN: An adaptive method for constructing high-resolution ocean models

Given the low resolution of ocean data, this paper proposes a method to construct a high-resolution ocean data model, adaptively selecting the optimal parameters for the model. The KNN regression model proposed in this paper is used to refine the seawater thermocline data, which significantly improves the data resolution on the vertical gradient. The proposed method is an improvement on the KNN regression model so that the model can adaptively select parameters based on the input data. The trained model can predict the temperature and salinity data at any position in the experimental area, and solve the low resolution caused by the sparse deployment of underwater nodes. The simulation results show that the proposed method can effectively select the optimal parameters for the KNN model.