摘要：As an open and complex giant system, the development factors of regional economy restrict and promote each other and finally bring a high degree of nonlinearity to various data reflecting regional economic development. In addition, due to the difficulty of statistics, economic data has the characteristics of less data and large errors. This article is aimed at studying a regional economic development forecasting algorithm in order to change the traditional forecasting modeling technology that is increasingly difficult to meet the needs of regional economic development forecasting. The management department can analyze the results of the data mining model to improve the administrative management of the management department, later supervision ability, and work efficiency. In this article, a regional linear regression prediction algorithm based on linear regression is established to support regression vectors. In order to ensure the reliability of the prediction results, the concept of prediction reliability is introduced into the LPSVR algorithm and specific calculation methods are provided. At the same time, it shows that the LPSVR forecasting model outperforms the neural network algorithm in the research of regional economic forecasting. The experimental results of this paper show that the prediction accuracy of Guilin’s annual GDP used by BP neural network is basically lower than that of LPSVR, which is 1.68%. LPSVR has shown good prediction performance in regional economic development prediction.
