Parking Generating Rate Prediction Method Based on Grey Correlation Analysis and SSA-GRNN

Abstract

The parking generating rate model is commonly used in parking demand forecasting. However, the key indicators of the parking generating rate are generally difficult to determine, especially its future annual value. The parking generating rate is affected by many factors. In order to more accurately predict the urban parking generating rate, this paper establishes a parking generating rate prediction model based on grey correlation analysis and a generalized regression neural network (GRNN) optimized by a sparrow search algorithm (SSA). Gross domestic product (GDP), urban area, urban population, motor vehicle ownership, and land use type are selected as input variables of the GRNN via grey correlation analysis. The SSA is used to optimize network weights and thresholds, and a model based on the SSA to optimize the GRNN is constructed to predict the parking generating rate of different cities. The results show that, after SSA optimization, the maximum absolute error of the GRNN model in predicting the parking generating rate is reduced, and the prediction accuracy of the model is effectively improved. This model can provide technical support for solving urban parking problems.