Hydrological time series prediction optimization method based on woa-lstm-mc

Abstract

The invention discloses a hydrological time series prediction and optimization method based on WOA-LSTM-MC, comprising using the optimized whale optimization algorithm to optimize some parameters of the prediction model; selecting the flow data of the hydrological station to be predicted as experimental data; The data set is divided into training set and test set for training and prediction; Markov chain MC is used for correction to obtain the final hydrological prediction result, that is, a more accurate prediction value; a hybrid WOA‑LSTM‑MC model is established. The invention can find the optimal parameters required by the prediction model faster and more accurately, and can ensure the global search ability and local exploration ability of the algorithm, the convergence speed is fast, and the prediction result is accurate.

Description

technical field [0001] The invention belongs to the hydrology prediction technology, in particular to a hydrology time series prediction optimization method based on WOA-LSTM-MC. Background technique [0002] With the development of today's computer and software technology, its application fields are becoming larger and larger, and massive time series data can be seen everywhere. On the one hand, some of these data have valuable intelligence for us to study. Data mining technology was born, and we can deduce the state of future data through the laws and development trends of the data itself, so it is very important for the analysis and prediction of time series. [0003] Hydrological data are discrete records of hydrological processes. Evaporation, rainfall, water levels, etc. are standard time series. Hydrological data has the characteristics of large data volume, fast update speed, and complex types. At the same time, it is also related to many conditions such as seasons,...

AI Technical Summary

Problems solved by technology

The effect of hydrological time series prediction is remarkable, but there are still some problems. For example, the accuracy of the model for small sample sets will be affected to a certain extent. There are problems that the larger the amount of data, the larger the error, and the more time-consuming the training. The prediction model itself is also difficult. There are some problems, so it is crucial to choose an algorithmic model with a certain robustness

Images