For instructions on how to create projects, annotate data, train models, and test models, see the Text Classification tutorial.
If you do not want to train a model, you can call an API service that uses a pre-trained model. For more information, see the following documents:
Interpreting training metrics
Overall accuracy: The accuracy of the model's predictions for all labels. This includes all property dimensions and all sentiment polarities, such as positive, negative, neutral, and not mentioned.
Category name=all: The aggregated metrics for all property dimensions.
Category name=all, Class=positive: The aggregated metrics for all positive sentiments.
Class=neutral: Samples with neutral sentiment are rare. The model's metrics for this class are usually low and may be 0.
该文章对您有帮助吗？