The Application Of Fuzzy Cluster Analysis In Extreme Temperatures

In daily life, production, scientific research work, often needs to classify subjects. The object of study in the old taxonomy, classification mainly depends on the experience and expertise to achieve. But with the development of science and technology, only depending on the experience and professional knowledge to classification has been unable to meet the needs of the people, so use mathematical methods to classify the cluster analysis and cluster analysis with either this or that character. But in real life, most of the research objects are also vital nature, no clear division between the classes of many things, the relationship between them is more of a fuzzy relation. This time, more needs to be done is a soft division. Fuzzy set theory for this kind of soft partitioning offers a powerful analysis tool, people began to use fuzzy method to deal with the problem of clustering, and the application of fuzzy mathematics method in cluster analysis, and fuzzy cluster analysis is. Fuzzy clustering has been widely used in data mining, image processing and other aspects. But there is no use in the extreme temperature.Weather condition analysis is a multifactor analysis. And extreme temperature is determined by the threshold. Thus, the ability to find a single factor fuzzy clustering to close to the traditional multi-factor of fuzzy clustering, it is the issue that article place should discuss. The basis of the theory of cluster analysis and fuzzy clustering analysis based on the traditional classification methods based on many factors, the proposed new factor classification, fuzzy clustering analysis and extreme weather. According to the20provincial capitals across the country from1965-200440years of data, extract the corresponding to the extreme temperatures, and with different thresholds of extreme temperature for different factors, finally using Suffer software produced by the Chinese base map and local data cluster analysis the results of comparative analysis, found the two clustering results with80%similarity. It also can be seen from the final results of the clustering in the country’s provincial capital, Xining’s weather conditions is the most special, Wuhan, Hangzhou, Nanjing, Hefei, Zhengzhou and other cities of the weather conditions is close to. Then, this article proposed classification method applied in Wuhan city50years of weather, get the corresponding classification, the best determine the threshold obtained when the Wuhan is divided into14categories is the most appropriate. It can be seen that Wuhan has33normal, and others abnormalous years.. According to historical data, abnormal years and historical data presentation data matches, once again proved the reasonability of the classification.