This is a binary classification model fine-tuned using the model 'bert-base-uncased'. It is built using a large Twitter dataset and is suitable especially for Twitter style data.
This can be used to classify the text into the categories of 'Privacy & Security' or 'Non-Privacy and Security'.
It achieved the following results on the evaluation set:
The validation scores for the module were as follows
Accuracy = 0.92
| Class | Precision | Recall | F1-Score |
|---|---|---|---|
| PrivSec(0) | 0.91 | 0.94 | 0.92 |
| Non-PrivSec(1) | 0.93 | 0.89 | 0.91 |
Paper: The paper detailing how it was designed can be found here Perspectives of non-expert users on cyber security and privacy: An analysis of online discussions on twitter
Please cite the paper if you use this model :
Nandita Pattnaik, Shujun Li, and Jason R.C. Nurse. 2023.
Perspectives of non-expert users on cyber security and
privacy: An analysis of online discussions on Twitter.
Computers & Security 125 (2023), 103008. https://doi.org/10.1016/j.cose.2022.103008
- Downloads last month
- 7