Improved CNN-GRU algorithm application in enterprise legal consultation system
With the rapid development of artificial intelligence technology in enterprise services, the application of intelligent legal consultation systems is becoming increasingly widespread. However, legal speech data have complex characteristics such as non-standard expression forms, diverse dialects, and...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Systems and Soft Computing |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772941925001693 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | With the rapid development of artificial intelligence technology in enterprise services, the application of intelligent legal consultation systems is becoming increasingly widespread. However, legal speech data have complex characteristics such as non-standard expression forms, diverse dialects, and highly specialized terminology. Existing speech recognition models still have obvious shortcomings in terms of accuracy and stability, making it difficult to meet the needs of practical applications. To this end, this paper proposes a hybrid model based on a deformable convolutional neural network structure and a gated recurrent mechanism. The study proposes an improved neural network that combines deformable convolution and gated recurrent units for convolution problems. By utilizing a deformable convolution mechanism, the model's ability to recognize non-standardized features of legitimate speech data has been improved. The time information of the voice data is captured by the gating loop unit. The results showed that the recognition accuracy of the model on the test set reached 98.17%, which was significantly better than AlexNet, bidirectional long short-term memory, and traditional gated recurrent unit. The performance of the receiver operating characteristic curve was also better. Although the calculation complexity was high, the average calculation time was 4.87 s. However, its excellent recognition performance proved its feasibility in high-precision enterprise legal consultation systems. The research model significantly improves the recognition accuracy and robustness of legal advice voice data. This model enhances the ability to understand complex legal terms and contexts. This paper provides new ideas and methods for improving the performance of intelligent question-answering systems. |
|---|---|
| ISSN: | 2772-9419 |