The area prediction of western North Pacific Subtropical High in summer based on Gaussian Naive Bayes

The western North Pacific Subtropical High (WNPSH) is a key circulation system regulating the East Asian climate, and its area has crucial indicative for summer precipitation in China. In this study, we established the model for classification prediction of summer WNPSH area via the Gaussian Naive Bayes (GNB). By setting different category proportions and different training set sample sizes, we investigated the prediction ability of GNB and its dependent on the data sample size. After comparing the prediction performance of GNB with tree models which were commonly used in short-term climate prediction, it was found that the accuracy scores (ACC) and balanced accuracy scores (BCC) of GNB were statistically significantly higher than tree models. Additionally, under different category classification criteria, the ACC and BCC of GNB could maintain above 0.77 and 0.75, respectively. Especially for anomalous categories, the recalls values could maintain above 0.5. These results indicate that the GNB had very strong prediction ability for the summer WNPSH area and could also better predict the degree of anomalous WNPSH area. Moreover, under different training set sample sizes, the ACC of GNB could be maintained above 0.6, which suggests that the GNB was less dependent on the data sample size and could reduce the limitation of abrupt interdecadal changes in climate on the available data sample size to some extent. This study reveals strong prediction ability of GNB for the summer WNPSH area, which also has high reference value for the research of other short-term climate prediction problems.