Application Of Time Series Models In Economic Data Analysis

Gross Domestic Product(GDP)is the core index of accounting for national economic activity.It is the market value of all final products(goods and services)produced by production factors in the economic society(a country or region)during a certain period of time.It is an important embodiment of the comprehensive economy of the area where the data belongs.Relatively accurate prediction of GDP is a practical issue in the economic field.Based on time series analysis,this paper establishes time series models with support vector regression(SVR)algorithm.Models are established through the economic situation similar to Jilin province and Heilongjiang province in 1952-2011,using SPSS to establish a simple linear regression model,the time series analysis and Eviews software to establish a simple index model and ARIMA model,and SVR to establish a support vector regression model with R language software.The establishment of two provinces' models are for the recursive prediction of GDP in 2012-2015,then calculate the mean square error using the four years' forecast value and known actual GDP data.The most suitable for the GDP model is the minimum mean square error.The optimal GDP model of Jilin province is ARIMA(1,1,0),and the optimal GDP model of Heilongjiang province is the support vector regression model.GDP is not only an index of economic progress in one province,but also an index of the economic situation and future possibilities between two provinces.Looking back 64 years,Heilongjiang province has been greater than Jilin Province in GDP.This paper selects the optimal application of two provincial model to forecast GDP during 2016-2020,then sees their future development trends and possibilities,and uses the difference between the two provinces to establish models to observe the gap on GDP and compare the changes in the rate of economic growth.Then based on Co-integration theory,it confirms that all of the industry and GDP in Jilin province have long-term co-integration relationships.On this basis,the error correction model is established that observes to be negative feedback mechanisms and the first industry has the greatest impact on Jilin GDP.At last,it analyzes the Granger causality test of them at different lag periods,and selects the optimal lag order.In this paper,different models are established in different methods,and the recursive prediction is used to choose the best model.The overall and partial analysis of the data provides valuable information and basis,for that the modeling and prediction of GDP in two provinces and the economic development of the internal industry in Jilin province.