Cloud model method for canonical correlation analysis of large data

Cloud model method for canonical correlation analysis of large data

Yang Jingli Wenping Zhang Jianji

In view of the present situation that the high complexity of the traditional large data canonical correlation analysis (CCA, canonical correlation analytical method) is no longer adaptable to the large data PB-level data scale, a large data CCA method based on cloud model is proposed. Based on the cloud computing architecture, this method through cloud operation, each endpoint cloud is merged into the central cloud, and then the central cloud droplets are generated, and the small sample is reconstructed by the central cloud droplet as the uncertainty of large data, and the low data quantity of the central cloud drops improves the computational efficiency. The experimental results on the real dataset verify the effectiveness of the proposed method.

Cloud model method for canonical correlation analysis of large data