数据挖掘技术在气象预测中的应用

发布时间：2018-03-08 12:14

本文选题：数据挖掘　切入点：关联规则　出处：《天津工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着全球天气的不断变化,依靠天气预报来及时发现灾害天气的出现显得尤为重要。大数据时代的来临,将数据挖掘技术应用于气象预报中,分析各种气象因子之间的关联,提高气象预报的准确性,具有十分重要的现实意义。传统的气象预报是基于统计的预测模型,采用概率领域的相关方法将历史数据建立一个或多个模型。但是,传统的统计方法往往适用于大量预报对象,预报对象越多,找出的预报因子和预报对象之间的关联越多,得到的统计结果越精确。然而,在实际应用中,往往需要针对某一特定天气对象进行预报,传统的统计方法存在一定的局限性。使用数据挖掘技术可以针对某一特定天气对象,快速处理海量天气数据,挖掘出潜在的、人们不易发觉的预报因子之间的关联,有助于提高天气预报的准确性。目前智能算法在数据挖掘领域的应用受到越来越多学者的关注。引力移动算法GMA是近年提出的一种启发式群体优化算法,性能比传统粒子群算法有着很大提高,但仍然存在着缺陷。为提高引力移动算法搜索性能,针对引力移动算法解决一些高维空间优化问题时存在的收敛速度慢、搜索精度不高的问题,本文提出一种基于亲和度的改进引力移动算法PGMA,即基于引力移动算法原理,通过构造一个基于亲和度概念的系数,对种群个体受到的引力合力公式作适当的变换来改造基本引力移动算法。改进后的算法对种群中个体的位置更新方向加以引导,提高算法的搜索精度和算法搜索能力。用13个基准函数对改进算法进行试验,验证了改进算法在求解精度和稳定性上优于基本引力移动算法。然后将PGMA算法应用到了关联规则挖掘领域,并通过实验证明其性能在关联规则挖掘领域中的提高,并将对这种关联规则挖掘方案进行独立性检验改进并应用到气象预测领域中。通过上海市气象数据集证明了具有独立性检验的TI-PGMA关联规则挖掘方案的准确性和有效性。
[Abstract]:With the continuous changes of the global weather, it is very important to rely on the weather forecast to discover the occurrence of the disaster weather in time. With the advent of big data era, the data mining technology is applied to the weather forecast, and the correlation between various meteorological factors is analyzed. It is of great practical significance to improve the accuracy of meteorological forecast. The traditional forecasting model is based on statistics, and the historical data are built into one or more models by using the method of probability domain. Traditional statistical methods are often applicable to a large number of forecasting objects. The more forecasting objects, the more correlation between the prediction factors and the prediction objects, the more accurate the statistical results are. However, in practical applications, It is often necessary to forecast a particular weather object, but the traditional statistical method has some limitations. Using data mining technology, we can quickly process the massive weather data for a particular weather object, and find out the potential weather data. The link between predictors that people are not easily aware of, At present, more and more scholars pay attention to the application of intelligent algorithm in the field of data mining. The gravitational movement algorithm (GMA) is a heuristic group optimization algorithm proposed in recent years. In order to improve the search performance of gravity moving algorithm, the convergence rate of gravity moving algorithm to solve some high-dimensional spatial optimization problems is slow. In this paper, an improved gravitational mobility algorithm based on affinity, PGMA, is proposed, which is based on the principle of gravitational mobility, and a coefficient based on the concept of affinity is constructed. An appropriate transformation of the formula of gravitational resultant force to the population individual is made to transform the basic gravity moving algorithm. The improved algorithm guides the orientation of the updating of the individual's position in the population. Improve the search accuracy and search ability of the algorithm. The improved algorithm is tested with 13 benchmark functions. The improved algorithm is superior to the basic gravity moving algorithm in solving accuracy and stability. Then the PGMA algorithm is applied to the field of association rule mining, and the performance of the improved algorithm in association rule mining field is proved by experiments. This paper improves the independence test of this association rule mining scheme and applies it to the field of meteorological prediction. The accuracy and validity of the TI-PGMA association rule mining scheme with independence test are proved by using the Shanghai Meteorological data set.
【学位授予单位】：天津工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P409;TP311.13

【参考文献】