Research on Quality Control of Marine Monitoring Data Based on Extreme Learning Machine

Abstract

The collection and research of marine monitoring data are the main work of marine science. How to obtain the observation data of un-known sea area is of great significance to marine scientific research. Using the data provided by the Argo project, in this work we aim to effectively predict the marine monitoring data with the extreme learning machine which is simple in structure with short training time, in order to achieve the accuracy of marine information. According to the size of the sample for adjusting the corresponding model parameters, the pre-diction of marine data can be achieved. By comparing the forecast data of the unknown sea area and the data originally provided by the Argo plan, the accuracy of the ocean data is higher and the trend of the reflected data is more detailed.