Ship flow prediction method based on deep learning

Abstract

The invention discloses a ship flow prediction method based on deep learning. The ship flow prediction method is based on an improved isolated forest-complementary set empirical mode decomposition-long short-term memory neural network. Firstly, the problems of noise and abnormal points in original data are considered, and the abnormal points in the data are eliminated by using an isolated forest algorithm; secondly, in order to further improve the prediction precision, input data is decomposed into intrinsic mode function components and residual components of different frequencies by using animproved complementary set empirical mode decomposition algorithm, each intrinsic mode function and residual error are predicted by using a long-term and short-term memory neural network separately, and finally, superposition reconstruction is carried out on a prediction result. According to the method, the prediction precision is improved, and the method has good adaptability to long-term or short-term time series data.

Description

technical field [0001] The present invention relates to the technical field of time series forecasting, in particular to a deep learning-based ship flow forecasting method, in particular to an improved IF-CEEMD-LSTM (isolated forest-complementary set empirical mode decomposition-long short-term memory neural network ) ship flow forecasting method. Background technique [0002] With the increase of modern maritime trade, the prediction accuracy of maritime ship flow is necessary to ensure the efficiency and safety of ship traffic, to broaden the theory of water transportation, and to port and maritime departments for port construction and development and effective use of waterways. However, the traffic flow forecasting process is more complicated and affected by many factors, such as season, GDP, port container throughput, port cargo throughput, etc. [0003] In the past, most of the traditional machine learning methods were used to model and predict ship flow. According to ...

AI Technical Summary

Problems solved by technology

Since a single model cannot satisfy the forecasting accuracy, current time series forecasting methods are an effective combination of two or more models

[0005] In view of the low precision of the existing time series forecasting technology, the characteristics of nonlinearity and non-stationarity of ship flow data, and the disappearance of gradients in the cyclic neural network, it is necessary to develop an improved IF-CEEMD-LSTM (IF refers to Isolation Forest, that is, Isolation forest algorithm; CEEMD refers to the improved complementary set empirical mode decomposition; LSTM refers to the Long Short Term network (Long Short Term Memory Neural Network) prediction method, which has good adaptability to long-term or short-term time series data prediction. Superior performance in prediction

Images