Inflation Rate Prediction Model Combining SMC-ABC Algorithm and BVAR

The inflation rate is an important indicator of the economic health of a country or region. To improve the accuracy of inflation rate prediction and to fill the shortcomings of the existing inflation rate forecasting model in terms of forecasting accuracy and adaptability, the study constructed a prediction model based on the technique of Bayesian vector auto-regression model. The experimental results revealed that the study’s improved Bayesian vector auto-regression model had a higher prediction accuracy than both the baseline and existing advanced models. Its mean absolute error and root mean square error were less than 0.001 and 0.015, respectively. In the cross-validation analysis, the generalization ability of the model was strong, with R-squared taking values of 0.962 and 0.984 for the test set and training set, respectively. In the goodness-of-fit test, the model had a significant advantage over the other models, with both Akaike information criterion and Bayesian information criterion in the minimum value interval of 0.1-0.3. Anticipation Likelihood Probability converged to a maximum value of 0.973, which was the most efficient in prediction. In the applied analysis, the method was confirmed by the Diebold-Mariano test to show predictive test P-values of 2.464 per cent and 3.486 per cent, respectively. The percentage deviation analysis of the sample’s percent bias had a small error, with all predictions being closest to the true value. The change trend and details of the prediction results were better grasped. The study scientifically and rationally aids policy formulation, helps to avoid the impact of inflation on the economic market, and promotes the stable and healthy development of the economy.