Research on Data Stream Frequent Itemsets Mining
|School||Beijing Jiaotong University|
|Keywords||Data stream Data Mining Data Stream Mining Frequent itemsets|
The data stream is a time- arrival of the set of items in an orderly way . And traditional static database data is different, the data stream is continuous , unlimited, usually to the arrival of the high speed and the data distribution is changed with time . Makes traditional frequent itemset mining algorithms are difficult to apply due to the characteristics of the data stream . Many researchers studied data stream mining frequent item sets . Currently, frequent itemsets mining data stream has become one of the basic problems in data mining . The characteristics of the data stream , the research and papers on key issues in data stream processing techniques and data stream mining . Research on some of the key issues to resolve technical papers . Papers on classic frequent itemset mining algorithms are introduced and experiments . Unlimited data stream can be seen through experiments and analysis of high - speed makes the classic frequent itemset mining algorithms are difficult to apply to the data stream . In addition, the paper introduced frequent itemsets algorithm for the current existing data flow analysis and summary. Finally the algorithms FP-CountMin . The algorithm to the data stream segments and take advantage of improved FP-growth algorithm for mining frequent sub- itemsets . Then , the Count Min Sketch itemset counting . Algorithm to solve the problem of fast and efficient compression statistical and computational . Through the experimental comparison algorithm and FP-DS , FP-CountMin algorithm has better time efficiency .