Application of Support Vector Machines and Holt-winters Model in Local Finance Forecast

摘要

Scientific fiscal policy is an indispensable element for achieving rapid regional development. Local finance, as an important part of national finance, is also the main reference in the formulation of fiscal policy, and is crucial to the impact of fiscal policy. In this paper, we choose the fiscal data of Guangxi as the representative, and use Holt-winters model and support vector machine model to forecast the fiscal revenue of relationship 2018–2019. Meanwhile, the performance of the two models on the problem is evaluated using MAPE. The study achieves more accurate forecast for local finance, which provides some help for the government to make fiscal policy and also offers some references for achieving sound national economic development.