Multi-dimensional uncertain data stream clustering algorithm

At present, there are some problems in the study of existing uncertain data stream clustering methods, such as the clustering model is apt to mismatch the data model of the uncertain data stream, and these methods usually assume that the probability density function, probability distribution function or probability of the uncertain data are known; however in real application system, the above information is hard to get. To solve these problems, a multi-dimensional uncertain data stream clustering algorithm, UIDMicro (Uncertain Interval Data Micro) based on interval data is proposed. In this algorithm, firstly, the interval data combining with the statistic information of uncertain data is used to represent the multi-dimensional uncertain data stream; then two levels of cluster windows, namely current cluster and candidate cluster are used to cluster the multi-dimensional uncertain data stream. Through adjusting the two levels of cluster windows dynamically, the real time matching of the clustering model and data model of the uncertain data stream is realized. The experiment results show that the proposed clustering algorithm possesses better clustering precision and higher processing efficiency.