为改变企业对未来销售量预测手段的单一性和随意性,设计开发基于.NET组件化的销售预测系统,从而给企业决策者一个有效的、简单明了的预测结果以指导决策。文中根据不同预测方法对某一组具体历史数据的敏感程度不同,通过.NET组件技术实现多种预测方法和预测精度的检验,自动遴选出最适合某一组数据的预测模型或预测方法,达到最佳的预测效果。最后,运用该系统对某汽车制造企业的销售量进行预测,得到较好的效果。
In order to avoid the oneness and random in selecting the means of forecasting sale quantity of an enterprise in future, a sales forecasting system is designed arid developed based on module technique on. NET platform, thus supplying the enterprise' s leader and decision maker with a valid, simple and manifest forecasting result for their decision- making. In view of the sensitivity of different forecasting means to the practical history data, various forecasting methods are implemented by using the Microsoft. NET module technique and the forecasting model or forecasting method (which is the most suitable to the history data) is automatically selected through the examination of the forecasting accuracy,so that can obtain the best forecasting result. At last, the forecasting system is used to forecast the sale quantity in an automobile manufacturing enterprise, which obtains a satisfying result.