As can be seen from the line graph above, when the word vector dimension is 100,
the effect is the best, the accuracy rate is 90%, the recall rate is 92.55%, and the F1
value is 91.26%. When the dimension is too high, the generalization ability of the
model to the feature data is reduced, resulting in a decline in model performance.
4.4 Comparison with Other Methods
In order to verify the validity of the model, this paper will compare with the more
advanced models; the results are shown in Table 3.
In this paper, the RCNN is used. From Table 3, the model has higher values in
accuracy, recall, and F1 than other models, which proves that the model used in this
paper has a good recognition effect.
Fig. 5. Changes in the results of different word vector dimensions
Table 3. Experimental comparison results of different models
Model
Accuracy (%) Recall (%) F1 (%)
Bi-LSTM 78.6
76.63
77.6
CNN
86.5
88.42
87.45
RCNN
90
92.55
91.26
An Incident Identification Method Based on Improved RCNN
103