Method and device for establishing timing sequence prediction model

Abstract

The invention belongs to the technical field of computers, and provides a method and device for establishing a timing sequence prediction mode. The method includes: acquiring K training sets from a plurality of samples, wherein the K is an integer greater than zero; acquiring an original feature set of each training set in the k training set; acquiring a fission feature set of each training set according to the original feature set of each training set; establishing a decision-making tree corresponding to each training set according to the fission feature set of each training set so as to acquire respective decision-making trees corresponding to the k training sets; generating a random forest according to the respective decision-making trees corresponding to the k training sets so as to establish a random forest based timing sequence prediction model. The method and device can improve the accuracy of the timing sequence prediction model.

Description

technical field [0001] The invention belongs to the technical field of computers, in particular to a method and device for establishing a time series prediction model. Background technique [0002] The time series forecasting model is based on the time series of historical statistics to predict and analyze the future trend of change. At present, the traditional time series forecasting models, such as the autoregressive integral moving average model, have certain requirements and assumptions for the data, have high requirements for the data quality, require white noise to have a normal distribution, etc., and usually assume a limited time window , using the data pattern (one-dimensional data) that occurred in the past to predict the future pattern, because the dimension is too low, the prediction information is too little, which leads to the low accuracy of the time series prediction model. Therefore, it is necessary to propose a new technical solution to solve the above tec...

AI Technical Summary

Problems solved by technology

At present, the traditional time series forecasting models, such as the autoregressive integral moving average model, have certain requirements and assumptions for the data, have high requirements for the data quality, require white noise to have a normal distribution, etc., and usually assume a limited time window , use the data pattern (one-dimensional data) that occurred in the past to predict the future pattern, because the dimension is too low, resulting in too little prediction information, resulting in low accuracy of the time series prediction model

Images