Weighted Fuzzy Markov Chains for Regions Research on Rainfall Forecasting
Xuan Huang, Xiaotong Liu, Chunfu Zhang · Journal of Physics Conference Series · 2023
Abstract The problem of rainfall prediction is of great importance for the study of the hydrological cycle of a watershed or region in China. Since rainfall series are not periodic as well as linear, rainfall series are measured based on Markov’s theorem, and the percentage of precipitation distance is used to classify the historical data. The traditional Markov chain was optimized and a weighted-fuzzy Markov rainfall prediction model was established. The results show that the average absolute error of the two prediction samples is 2.09%, except for 2019, when the rainfall in Guangzhou City is different from the past, and the model has high accuracy, which has good practical significance and extension value.