A TEMPORAL CONTINUITY CLUSTERING ALGORITHM FOR SPATIAL-TEMPORAL DATA BASED ON GMM AND KNN

With the development of technology and the further exploration of space, the acquisition of numerous spatial-temporal data has become possible. How to use data mining technology, especially use clustering method, to analyze and derive information from spatial-temporal data to more effectively cope with real-world problems, has become an important issue. Based on the spatial-temporal temperature dataset covering 42 latitude and longitude points in one year, a new clustering algorithm combining GMM and KNN is developed to solve the problem of non-continuous time for within-cluster points. Then, this paper applies this clustering algorithm to the datasets of other time points in the same year and to the datasets of the subsequent year to explore the transferability of the algorithm and the availability of the classification results. The application results show that the proposed clustering algorithm can achieve temporal clustering of spatial-temporal data based on temperature while ensuring the temporal continuity of within-cluster points and the availability of classification results.