- 收稿日期: 2017-11-14
- 录用日期: 2018-03-12
Large Data Mining Method Based on Semantic Correlation Feature Fusion
- Received Date: 2017-11-14
- Accepted Date: 2018-03-12
Abstract: A large data mining algorithm based on semantic correlation feature fusion is proposed. Phase space reconstruction of the cloud storage large distributed data flow is taken for information extraction, the semantic association feature is extracted in the reconstruction phase space, the extracted features are taken as the testing sets for the adaptive training. The fuzzy C means algorithm is taken for the big data semantic correlation feature sparsity fusion and the clustering processing, the directional clustering of mining target data is realized in the cluster center, the mining data is output, and the feature compressor is used to reduce the dimension and reduce computational overhead. Simulation results show that the method can mine the big data accurately, the clustering ability is stronger, and it has the advantages in real-time and accuracy.