气象数据深度挖掘优化方法研究与仿真  

Depth of Meteorological Data Mining Research and Simulation Optimization Method

作　　者：李雷孝[1] 

机构地区：[1]内蒙古工业大学信息工程学院,内蒙古呼和浩特010051

出　　处：《计算机仿真》2013年第12期403-406,共4页Computer Simulation

摘　　要：研究气象数据准确预报问题。气象预报的相关数据属性包含众多特征,呈现多层次特性,数据之间需要运用一套极其复杂的数学方程来描述大气的运动规律。传统挖掘方法无法考虑众多气象因素的多层次性,仅仅根据数据浅层关系推断大气特征,更难以分析数据属性间隐含的信息,造成数据分析效果不理想。为了避免上述传统算法的弊端,提出了一种基于人工免疫算法的气象数据深度挖掘方法。利用聚类算法,对气象数据的特征进行准确的提取,从而为气象数据深度挖掘提供依据,利用聚类方法,能够对气象数据特征进行准确的筛选,在每个气象数据组中选择一个具有较强代表性的数据,并且将其余数据删除,降低气象数据的冗余性,并对剩余的数据通过交叉变换的方法进行特征筛选。实验结果表明,利用改进算法能够提高气象预报的精度,为人们的出行和工业生产提供保障。In this paper, the accurate prediction problem of meteorological data was studied. The property of weather forecast data contains numerous features and it presents multi - level characteristics. Highly complex mathematical equations have to be used to describe the movement of the atmosphere among these data. Traditional mining methods can not consider the multi level nature of many meteorological factors. And based solely on the shallow re lationships of data to extrapolate atmospheric characteristics, it is more difficult to analyze the implicit messages between data attributes, which results in a non ideal data analysis. In order to avoid the above drawbacks of traditional algorithm, we proposed a new depth meteorological data mining method based on artificial immune algorithm. To use clustering algorithm, the characteristics of meteorological data were accurately extracted, so as to provide the basis for depth mining of meteorological data; and it can accurate screen the meteorological data characteristics. In each group of meteorological data, more strong representative data were selected, the remaining data can be deleted to reduce the redundancy of meteorological data, and the feature selection of remaining data was used by cross - transform method. Experimental results show that the algorithm can improve the accuracy of weather forecasting, and provide the protection for travelling and industrial production.

关 键 词：气象数据 挖掘 大气运动 人工免疫 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]