NLP4CyberSecurity
NLP4CyberSecurity copied to clipboard
NLP model and tech for cyber security tasks
NLP4CyberSecurity
中文说明 | English
This code is NLP models and tech implementation for cyber security task, driven by deep learning model, a nice work on cyber security.
本项目使用自然语言处理(NLP)技术应用于网络安全领域,包括恶意软件检测、漏洞发现和威胁情报等方面。该项目基于Python编程语言和机器学习框架Scikit-learn、TensorFlow和Keras等,实现了一些常见的NLP技术,如文本预处理、特征提取、词嵌入、文本分类和主题建模等。通过对网络安全方面的文本数据进行处理和分析,该项目能够提高网络安全人员的工作效率和准确性,以及更好地发现网络安全威胁。此外,该项目还提供了一些用于网络安全的NLP数据集和预训练模型,方便其他研究人员和开发者使用。
- [x] Dataset
- [x] weak password
- [x] xss injection
- [x] malicious url
- [x] phishing url
- [x] Usage
- [x] Training
- [x] Example
- [ ] Demo
- [x] Reference
Update
Table of Contents
-
NLP4CyberSecurity
- Update
- Table of Contents
- Requirement
- Usage
-
Weak Password Detection
- Eval Result
-
XSS Injection Detection
- simple nn model
- simple cnn model
- simple lstm model
-
Malicious URL Detection
- RNN
- CNN
- Conv LSTM
- Phishing URL Detection
- Demo
- Star-History
- Reference
- Donation
- License
Requirement
pip install -r requirements.txt
Usage
Weak Password Detection
weak password detection with machine learning
weak-password/password-strength detection with machine learning; 弱密码检测;密码强度检测
Eval Result
precision recall f1-score support
0 0.94406 0.83240 0.88472 8920
1 0.96327 0.98971 0.97631 49652
2 0.99035 0.95400 0.97184 8392
accuracy 0.96428 66964
macro avg 0.96589 0.92537 0.94429 66964
weighted avg 0.96410 0.96428 0.96355 66964
XSS Injection Detection
xss injection detection with machine learning
simple nn model
Precision score is : 0.9764296754250387
Recall score is : 0.9830772223302859
simple cnn model
Precision score is : 0.9948463825569871
Recall score is : 0.9762692083252286
simple lstm model
Precision score is : 0.9980311084859225
Recall score is : 0.9869548286604362
Malicious URL Detection
malicious url detection with machine learning
RNN
Accuracy Score is: 0.8655441478439425
Precision Score is : 0.8579050828418984
Recall Score is : 0.8767578205075642
F1 Score: 0.8672290036092299
AUC Score: 0.8655252346603806
CNN
Accuracy Score is: 0.8379671457905544
Precision Score is : 0.8431494883953082
Recall Score is : 0.831085236357673
F1 Score: 0.8370738958974254
AUC Score: 0.8379787529437384
Conv LSTM
Accuracy Score is: 0.9242505133470226
Precision Score is : 0.9288969917958068
Recall Score is : 0.9191095076052642
F1 Score: 0.92397733127254
AUC Score: 0.9242591842604873
Phishing URL Detection
phishing url detection with machine learning
accuracy: 0.9982
Model Accuracy: 99.82%
precision recall f1-score support
0 0.99790 0.99895 0.99843 1904
1 0.99866 0.99732 0.99799 1495
accuracy 0.99823 3399
macro avg 0.99828 0.99814 0.99821 3399
weighted avg 0.99824 0.99823 0.99823 3399
Demo
Samples:
Star-History
Reference
Donation
If this project help you reduce time to develop, you can give me a cup of coffee :)
AliPay(支付宝)
![ali_pay](https://github.com/jackaduma/NLP4CyberSecurity/raw/main/misc/ali_pay.png)
WechatPay(微信)
![wechat_pay](https://github.com/jackaduma/NLP4CyberSecurity/raw/main/misc/wechat_pay.png)
License
MIT © Kun