Parallel frequent subgraph mining algorithm
| Autor: | He Yanshan, Xie Jianli, Zhang Ming, Wang Ting |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Theoretical computer science
Subgraph isomorphism problem Process (computing) 020206 networking & telecommunications 02 engineering and technology Set (abstract data type) Identification (information) 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Induced subgraph isomorphism problem Isomorphism Enhanced Data Rates for GSM Evolution Algorithm Time complexity MathematicsofComputing_DISCRETEMATHEMATICS Mathematics |
| Zdroj: | ICSCA |
| Popis: | This paper presents a new frequent subgraph parallel mining algorithm. The algorithm divides the whole process into two parts. Firstly, the master processor generates frequent subtree set by using existing subtree isomorphism judgment algorithm, and distributes the generated subtrees set to other slave processing nodes. Secondly, the slave processing nodes parallel process frequent subgraph edge expansion and isomorphism identification which have the highest time complex in frequent subgraph mining algorithm, so as to achieve purpose of improving the overall efficiency of the algorithm. The time complexity of the algorithm proposed in this paper isO(2n × n2/k) , in which n is the number of frequent edges and k is the number of the slave processing nodes. The experiment choose the open dataset and shows the new algorithm has good performance on efficiency and expansibility. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|