Time sequence prediction method and device, equipment and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a time sequence prediction method and device, equipment and a storage medium, and the method comprises the steps: obtaining a target time sequence; inputting the target time sequence into a target time sequence prediction model for time sequence prediction, wherein the target time sequence prediction model is obtained through training according to a generation module, a judgment module and a time sequence prediction initial model, and the time sequence prediction initial model is a module obtained according to any one of an LSTM network, a GRU network and an ARIMA model; and obtaining a time sequence prediction result corresponding to the target time sequence output by the target time sequence prediction model. The generation module and the judgment module are beneficial to quickly expanding training samples, so that the target time sequence prediction model obtained through final training has better robustness, and the success rate and accuracy of prediction of the target time sequence prediction model are improved.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, in particular to a time series prediction method, device, equipment and storage medium. Background technique [0002] With the development of machine learning technology at any time, machine learning technology is used to train time series forecasting models, and time series forecasting through trained time series forecasting models has been widely used. The time series prediction models in the prior art are all trained using historical time series. For example, in order to predict the 13-month continuation rate of institution A in the next month, the 13-month continuation rate in the history of institution A is required. Because of the historical time series data The amount of abnormal data is relatively small, and the abnormal data is not fully represented in the historical time series, which leads to poor robustness of the trained time series forecasting model, and ma...

AI Technical Summary

Problems solved by technology

[0003] The main purpose of this application is to provide a time series forecasting method, device, equipment and storage medium, aiming to solve the problem of using historical time series to train time series forecasting models in the prior art. Due to the relatively small amount of historical time series data, abnormal data Insufficient representation in the historical time series leads to poor robustness of the trained time series forecasting model. When the trained time series forecasting model faces new scenarios, it may fail to predict and / or have low accuracy. low technical issues

Images