2018 Volume 33 Issue 3 Pages G-SGAI05_1-11
Globally-covered ocean monitoring system Argo with more than 3,700 autonomous floats has been working, and its accumulated big ocean observation data helps many studies such as investigation into climate change mechanism. Since the observed data sometimes involves errors, human experts must visually confirm and revise quality control (QC) flags. However, such manual QC by human experts cannot be performed in some countries. In addition, it is difficult to regularize the quality of the ocean observation data of all over the world because the manual QC depends on human experts’ heuristics. Therefore, this paper proposes a method for error detection in Argo observation data using Conditional Random Field (CRF) because the problem requires consideration of sequence of both features and quality flags for accurate labeling in each depth. This paper also proposes a feature function design method using decision tree learning, allowing coping with various types of observation errors without manual work, whereas previous work had to focus on certain error types due to manual labor for feature function design. Furthermore, the proposed method divides the two CRF-based sequential classifiers that use manually- or automatically-designed feature functions respectively rather than combining the both feature functions into a single set. Experimental results have shown that the proposed method could detect all types of salinity errors with higher accuracy of QC flags assignments than the actually operated system in Argo project. In particular, the recall rate of the proposed method was better than that of CRF using the manually designed feature functions even for the specific error types for which they were designed.
