Interactive Metadata Fast Clustering Method Based on Improved Association Rules
Aiming at the problems of uneven distribution of interactive metadata samples,long clustering time and poor detec-tion effect of outliers,an interactive metadata fast clustering method based on improved association rules is proposed.The in-teractive metadata are divided into multiple data slices to form a metadata distributed processing framework.According to the i-dentification and positioning information of the data object to be clustered,the data grid index mechanism is established.The Boolean matrix and inner product algorithm are used to replace the coding method and operation method of the basic association rules,define the value function required by the classification algorithm,change the arrangement mode of feature points into sor-ting by data dimension,and obtain improved association rules.Combined with fuzzy mean algorithm and Lagrange multiplier method to achieve interactive metadata fast clustering.The experimental results show that compocred with the original meth-ods,the clustering running times of the proposed method are reduced by 81.27％and 79.03％,which effectively improves the efficiency of metadata clustering.
improved association ruleBoolean matrixdata clusteringinteractivemetadata
万方数据
维普
