Automobile Inventory Sales Prediction Model Based On Time Series Data Mining

Automobile industry is regard as underpinning of country economy, which has very important role in economic developments. Recently, automobile inventory industry is increasing with the high speed steps of automobile industry and continual extending scales. With the enhancements of informational and enrichments of sales data, it is necessary for managers to do forecast sales. Franchisers need to forecast market requirements of automobile inventory to maintenance appropriate storage, which must satisfy market demands. Consequently, how to find a good data mining methods is a vary importance and worthy research task.This thesis describes a sales forecast model with automobile inventory data. The data mining process starts with the character of historical sale data, establishes and optimizes forecasting models. This paper adopts radial basis function, back propagation neural network and differential evolution back propagation network to establishes three different forecast models, establish a hybrid forecasting model based on least squares analysis, which takes the advantages of those three models. The hybrid model appears to have better stability and precision when used to predict the given data.