论文部分内容阅读
时间序列预测模型是基于与预测对象历史数据相关的趋势、规律以及数据间外部联系而建立的数学算法模型,一般分为短周期预测和中长周期预测,预测模型是数据挖掘算法的一个分支,预测模型只是将预测对象原有的一些内在规律及与之相关的数据间的外部联系用数学算法加以发掘分析,而对于一些随机的即缺乏内在规律的或者关联性较弱的数据,现有预测模型一般难以达到理想的效果。本文详细介绍了时间序列预测模型预测理论及其约束,然后介绍该预测模型在炼油企业瓦斯气柜负荷预测中的实际应用,对生产进行预测及指导。
The time series forecasting model is a mathematical algorithm model based on the trends and regularities related to the historical data of the forecasting objects and the external relations between the data. Generally, the time series forecasting model is divided into the short cycle forecasting and the mid-long term forecasting. The forecasting model is a branch of the data mining algorithm, The prediction model only excludes and analyzes the external relations between the original internal laws of the forecast objects and the data related to them. However, for some random data, that is, lack of internal rules or weak correlations, the existing prediction The model is generally difficult to achieve the desired effect. This paper introduces the theory and constraints of time series forecasting model in detail, then introduces the practical application of this forecasting model in the forecasting of gas tank load of oil refining enterprises, and forecasts and guides the production.