Prediction method of regional economic development level based on SOM algorithm

As a vital but time-consuming task, regional GDP forecasting has high expectations for forecast accuracy in order to provide better guidance and suggestions for the region's economic development in the future.Most traditional GDP pre-methods are linear, but their accuracy has gradually been difficult to meet the demand due to factors such as nonlinearity and GDP uncertainty.This paper proposes a method for forecasting the level of regional economic development based on the SOM algorithm in order to improve forecast reliability and accuracy.These two-dimensional neural networks, with their layers of neurons arranged in a two-dimensional topology, are combined in this paper to create a self-organizing feature map model (SOM).The ANN model converts the SOM's output and outputs the model's final classification result directly.Using this model's algorithm, the model's performance can be greatly improved, while noise samples can be eliminated and the model's accuracy greatly improved.This study predicts future regional GDP using data on prefecture-level city GDP as the regression independent variable from 2001 to 2021.This paper's method successfully predicts the level of economic development in a region, as demonstrated by a large number of experiments.