**6. Conclusions**

In this work, we propose a multiple attention mechanism graph convolution action recognition model based on coordination theory (MA-CT). It parameterizes the graph structure of the skeleton data and embeds it into the network to be jointly learned and updated with the model. This data-driven approach increases the flexibility of the graph convolutional network and is more suitable for the action recognition task. Furthermore, the existing methods do not make full use of the coordination features of human motion, and because of the existence of an adjacency matrix, the model cannot extract features from the global perspective. In this work, we propose a coordination attention module (CAM) and importance attention module (IAM). In this paper, experiments are carried out on two large public datasets. For the two indicators of NTU-RGB + D, the CAM improves the accuracy of the model by 0.2% and 0.3%, and the IAM improves the accuracy of the model by 0.6% and 0.4%. In the Kinetics dataset, the CAM improves the accuracy of the model by 0.4%, and the IAM improves the accuracy of the model by 0.9%. They are used to solve the problems of insufficient feature extraction and the capturing of key joints. The final model has achieved good results in NTU-RGB + D and Kinetics.

**Author Contributions:** Conceptualization, K.H. and Y.D.; methodology, K.H. and H.H.; software, Y.D.; validation, Y.D.; formal analysis, Y.D.; investigation, Y.D. and J.J.; resources, K.H. and Y.D.; data curation, K.H.; writing—original draft preparation, Y.D. and M.X.; writing—review and editing, Y.D.; visualization, Y.D. and J.J.; supervision, K.H.; project administration, K.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** Research in this article is supported by the key special project of the National Key R&D Program (2018YFC1405703), and the financial support of Jiangsu Austin Optronics Technology Co., Ltd. is deeply appreciated.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable. **Data Availability Statement:** The data and code used to support the findings of this study are available from the corresponding author upon request (001600@nuist.edu.cn).

**Acknowledgments:** We would like to express my heartfelt thanks to the reviewers and editors who submitted valuable revisions to this article.

**Conflicts of Interest:** The authors declare no conflict of interest.
