Webshell Detection

Webshell Detection Based on the Word Attention Mechanism

Read this article

Cloned from https://github.com/leett1/Programe/

Editing directory project.zip

Install packages:

• gensim==3.8.1
• python-Levenshtein==0.12.0
• pathlib==1.0.1
• numpy==1.19.2
• tensorflow==1.14
• keras==2.3.1
• scikit-learn==0.24.1

Initializing a machine learning environment

Deploying a virtual environment Anaconda3

• Install packages

  yum install -y zlib-devel bzip2 bzip2-devel readline-devel sqlite sqlite-devel openssl-devel xz xz-devel libffi-devel

• Install Anaconda3

  curl -O https://repo.anaconda.com/archive/Anaconda3-2019.03-Linux-x86_64.sh

  bash Anaconda3-2019.03-Linux-x86_64.sh

• Edit file .bashrc

  vim ~/.bashrc

  Add this text to the end of the file:

  export PATH="/home//anaconda3/bin:$PATH"

  <user> - replace with user

• source ~/.bash_profile

• Install environment

  conda create -n tf1_env

  conda activate tf1_env

Project ML

• cd /var

• mkdir cnn_word2wec_sentence

• chmod 775 cnn_word2wec_sentence

• chown <user>:<group user> cnn_word2wec_sentence

• cd /var/cnn_word2wec_sentence

Remote Python interpreter

• Interpreter: /home/<user>/anaconda3/envs/tf1_env/bin/python

  <user> - replace with user

• Project migration: /var/cnn_word2wec_sentence

Python libraries

• Project path

  cd /var/cnn_word2wec_sentence

• Activate env

  conda activate tf1_env

• Install TensorFlow

  conda install tensorflow==1.14.0

  conda install keras==2.3.1

• Install requirements

  pip install -r requirements.txt

• Check

  python -V

  python -c 'import tensorflow as tf; print(tf.__version__)'

Train & test models

1. Model training: python3 one_attention_model.py

   Output: one_attention_mode190626_dan.h5

2. Model training: python3 train_model.py

   Output: two_attention_mode190317.h5

3. Model training: python3 word2vec_train.py

   Output: word_train190313.model

4. Edit the file and run the model test: python3 test_1.py