Multi-view Sparse Subspace Clustering And Its Application In Climate Zoning

Climate zoning is the division of specific regions based on some meteorological elements,and has important application value in the fields of agricultural production and building energy conservation.Most of the existing climate classification methods only consider one climate variable,but in practical applications,we need to consider the combined effects of multiple climate variables on the climate classification at the same time.Therefore,the study of climate classification is a very important subject.This thesis will study the application of multi-perspective clustering method in climate zoning.The main work is as follows.Part of the meteorological data from China's 661 meteorological stations are selected to test the correlation among relative humidity,atmospheric pressure,sunshine hours,daily average temperature and daily temperature difference,and the maximum information coefficient among them are calculated.The experimental results shows that the chosen meteorological elements do not have strong correlation,which is conducive to multi-view clustering.Multi-view low-rank sparse subspace clustering method is proposed to partition China's climate.A multi-view low-rank sparse subspace clustering model is first established,and the alternating direction multiplier method is employed to solve the model.Based on the linear kernel function and Gaussian kernel function,the selected meteorological stations are clustered from multiple perspectives.The clustering results are compared with the k-means clustering results,which proves the rationality of the proposed method.Statistical analysis is performed on air conditioning degree days,heating degree days,average temperature of the coldest month,and average temperature of the warmest month in each climate zone.Experimental results show that this method can effectively partition China's climate.A multi-view subspace clustering model combining feature weights and local structure is established,this method takes into account the global structure and local structure of each view simultaneously,and the alternate direction multiplier method is used to solve the model.Applying the above method to China's climate zones,statistical analysis is carried out on the air-conditioning degree days,heating degree days,the coldest month average temperature and the warmest month average temperature of each climate zone.The experimental results demonstrate that the divided climate zones have obvious boundaries,so this method can effectively divide the climate of China.