[Retracted] Research on Forecast of Macroeconomic Indicators Based on Multiobjective Optimization
Table 8
Correlation analysis of different algorithms.
Name
Correlation index set
Relative error
PSO-BP algorithm
Productive loans, industrial growth value, M2, raw materials, fuel; foreign exchange reserves, total import and export volume
6.13%
ARIMA model
6.30%
PSO-BP algorithm
Productive loans, industrial growth value, M2, raw materials, fuel, foreign exchange reserves, total import and export volume
0.67%
Algorithm based on kernel principal component and weighted SVM
Gross domestic product, residents’ consumption expenditure, government purchase, investment in fixed assets
4.13%
BP neural network
5.26%
Combination forecasting algorithm based on grey deep perceptron
Gross product, total investment in fixed assets of the whole society, total retail sales of social consumer goods, total tourism revenue, total import and export volume
2.99%
Multiple regression model and ridge regression model
Gross domestic product, fiscal expenditure, retail price index of commodities
4.53%
Elasticity coefficient method of energy consumption based on fixed base
Real annual GDP growth rate, the actual average annual growth rate of the added value of secondary production, average annual growth rate of employed population in secondary production, proportion of added value of tertiary industry
1.40%
Based on LASSO regression and GM (1, ) model
Effective irrigation area, application amount of chemical fertilizer, rural electricity consumption, total power of agricultural machinery, sown area of grain crops, grain yield per unit area
