Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Training from scratch.

More »

Fig 1 Expand

Fig 2.

Pre-training and fine-tuning.

More »

Fig 2 Expand

Fig 3.

Prompt learning.

More »

Fig 3 Expand

Fig 4.

EMSA prompt engineering.

More »

Fig 4 Expand

Table 1.

GubaSenti dataset statistics.

More »

Table 1 Expand

Table 2.

10 examples of GubaSenti dataset.

More »

Table 2 Expand

Fig 5.

An example of 1-shot standard prompt template.

More »

Fig 5 Expand

Table 3.

Comparison of pre-training and fine-tuning models and prompt engineering on GubaSenti sentiment analysis on the accuracy, precision, recall and F1-score (%).

More »

Table 3 Expand

Fig 6.

Prompt templates with varied combinations of English and Chinese.

More »

Fig 6 Expand

Table 4.

Comparison between 3 prompt templates on GubaSenti sentiment analysis on the accuracy, precision, recall, and F1-score (%).

More »

Table 4 Expand

Fig 7.

EMSA propmt example.

All of the content used to construct the template is generated by GPT-4 API and the correctness and effectiveness of the generated content have been verified manually. (The English translation is only an English version of the EMSA prompt template and it is not relevant to the experiment.)

More »

Fig 7 Expand

Table 5.

Comparison of different prompting approaches.

More »

Table 5 Expand

Table 6.

Overview of baseline models for sentiment analysis comparison.

More »

Table 6 Expand

Fig 8.

Accuracy of multiple GPT models on GubaSenti with different number of templates.

More »

Fig 8 Expand

Table 7.

Performance comparison on GubaSenti dataset.

More »

Table 7 Expand

Fig 9.

Confusion matrices for Gubasenti.

More »

Fig 9 Expand

Table 8.

Performance comparison on SST-2 dataset.

More »

Table 8 Expand

Table 9.

Performance comparison on SST-5 dataset.

More »

Table 9 Expand

Fig 10.

Confusion matrices for SST-2.

More »

Fig 10 Expand

Fig 11.

Confusion matrices for SST-5.

More »

Fig 11 Expand

Table 10.

Ablation Study Results on GubaSenti Dataset.

More »

Table 10 Expand

Table 11.

Ablation study results on SST-5 dataset.

More »

Table 11 Expand

Table 12.

Major error categories and their distribution.

More »

Table 12 Expand

Table 13.

Examples of successful implicit sentiment analysis.

More »

Table 13 Expand

Table 14.

Examples of successful context integration.

More »

Table 14 Expand

Table 15.

Examples of sarcasm-related errors.

More »

Table 15 Expand

Table 16.

Examples of mixed sentiment errors.

More »

Table 16 Expand