Indexed by:
Abstract:
In order to solve the problem that it is difficult to efficiently identify homologous complaints and reports in dealing with complaints and reports, this paper proposes to detect homologous complaint reports in the decomposed attention network. Firstly, TextCNN is used to extract local features of complaint reports text, and then the matching task between features is carried out. Due to the high noise of the complaint reports text, the matching method of a few key features is better than capturing the global features of the text when judging whether the two complaint reporting tasks are the same origin events. The resolvable attention network decomposed the same-origin complaint reporting task into the matching task between recognition channels, and then judged the relationship between complaint reporting according to the judgment results of each sub-task. This paper compares the accuracy of multiple neural network models on the complaint reporting data set, and the experimental results of these models show that the decomposition of the attention network model in this paper has a better effect and stronger stability. © 2022 Technical Committee on Control Theory, Chinese Association of Automation.
Keyword:
Reprint Author's Address:
Email:
Source :
ISSN: 1934-1768
Year: 2022
Volume: 2022-July
Page: 7197-7202
Language: English
Cited Count:
SCOPUS Cited Count:
ESI Highly Cited Papers on the List: 0 Unfold All
WanFang Cited Count:
Chinese Cited Count:
30 Days PV: 0
Affiliated Colleges: