Carrie rocket data prediction method and device, storage medium and electronic device

Abstract

The invention discloses a method and a device for predicting data of a carrier rocket, a storage medium and an electronic device. The method comprises the following steps: step S101, determining a weighting coefficient vector K; 102, calculating a weighted prediction value set ZP_final based on that weight coefficient vector K; Step S103: based on the comparison result of the weighted prediction value set ZP_final and the actual value of the rocket test data corresponding thereto, all weighted prediction values in the weighted prediction value set ZP_final are outputted. The present application combines the AR model as a stochastic model with the regression analysis model as a deterministic model, and is applied to the prediction of rocket test data, so as to not only measure the influenceof a certain factor, but also predict the upcoming data according to the already tested data.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a data prediction method and device for a launch vehicle, a storage medium and electronic equipment. Background technique [0002] In the 1970s, G.E.P.Box and G.M.Jenkins developed a discrete parameter prediction recursive algorithm (Box-Jenkins model) that satisfies the finite parameter linear model, which made it possible to combine prediction theory with computers. With the development of computer technology, models such as AR, MA, ARMA and ARIMA have been widely used. [0003] At present, research on combined forecasting models is a hot spot. Many scholars have proposed combined time series forecasting methods, such as the time series forecasting method based on Hilbert-Huang transform and ARMA model, the time series forecasting method based on wavelet analysis and rolling time series forecasting algorithm. Optimization forecasting methods, forecasting methods based ...

AI Technical Summary

Problems solved by technology

The most widely used is simple weighted average, but this method cannot use the information that contains the prediction accuracy before, so the method of weighted average has the problem of low accuracy

Another simple method is to linearly mix the independent forecasts, that is to say, combine a part of the data from the two matrices of the past forecast value and the past observation value into a variety of matrices, and find the correlation in the various matrices. The most reliable matrix, and then determine the combination weight by the method of least squares (Ordinary Least Square, OLS). Due to the large number of matrices, the amount of processing required is large, resulting in the inefficiency of the least squares method

For the trend prediction of launch vehicle test data, few people use the method of time series prediction

Images