Application Research Of Data Mining In Meteorological Data Analysis

The capacity and field of meteorological data are growing and expanding rapidly as time goes by, forming Data Mountains and Data Mazes. Workers who work on the field of meteorology have faced the problem that how to effectively use meteorological data. The traditional methods in computer sciences have great difficulty in dealing with data on a large scale. Therefore, data mining is taken into consideration.In this work, the research and application status of data mining at home and abroad in meteorological data analysis was firstly reviewed and the achievements and disadvantages of data mining in the analysis of meteorological data were elucidated. Secondly, multidimensional data set bank was established for meteorological data of each hour Baotou meteorological data in three consecutive years. Four data processing methods including data cleaning, data integration, data transformation and data reduction were utilized to preprocess data for meteorological multidimensional data set bank in order to improve the quality of data mining objects, achieving purposes of modes, knowledge, norms etc. for the improvements of data mining. Then, the application of four major data mining technologies in meteorological data analysis was studied. They are enumerated as follows: the utilization of isolated point analysis method is uesd to analyze exception record sets in the meteorological data banks, and explore some abnormal value and unusual weather record set , the rainfall prediction model and the pollution model were builded by use of decision tree which can dig such rules that the concentration of pollutants will exceed standard under any meteorological condition, the adoption of class methods to classify meteorological data to detect meteorological characteristics-proposing hierarchical clustering based on class profile which is applied on meteorological data to prove the feasibility, validity and accuracy of the algorithm, the exploitation of association rule mining to study related meteorological phenomena between meteorological. In the end, SPSS Clementine as a mainstream data mining software and some classical data mining model were put into use to dig data in the meteorological database of the city of Baotou.In this study, major theoretical and practical values of data mining mainstream technologies utilized in the analysis of meteorological data were discussed on the basis of the research work. At the same time, some latent and useful models, rules and knowledge were dug through the analysis of meteorological data of the city of Baotou.