Skip to content

Latest commit

 

History

History
 
 

Hayawi

DeeProBot: a hybrid deep neural network model for social bot detection based on user profile data


├── boto-19/
├── cresci-18/
├── cresci-17/
├── cresci-19/
├── twibot-22/
├── midterm-18/
├── cresci-15/
├── twibot-20/
├── gilani-17/
├── run.py 
└── Hayawi.md # README
  • implement details: “Sentiment”, “Timing” features are discarded since required information is not included in datasets.

How to reproduce:

  1. preprocess the dataset and train this model by running

    python run.py {dataset name}

    twibot-22 for example :

    python run.py twibot-22

    • dataset names:
      • boto-19
      • cresci-18
      • cresci-17
      • cresci-19
      • twibot-22
      • midterm-18
      • cresci-15
      • twibot-20
      • gilani-17

Result:

dataset acc precison recall f1
Cresci-2015 mean 0.8427 0.9296 0.7931 0.8556
Cresci-2015 std 0.0002 0.0003 0.0002 0.0001
Twibot-20 mean 0.7314 0.7161 0.8350 0.7705
Twibot-20 std 0.0001 0.0001 0.0004 0.0002
Twibot-22 mean 0.7650 0.8000 0.1499 0.2474
Twibot-22 std 0.0007 0.0027 0.0005 0.0008
Gilani-17 mean 0.5270 0.5144 0.2800 0.3467
Gilani-17 std 0.0002 0.0005 0.0013 0.0011
Cresci-2017 mean 0.9078 0.9547 0.9219 0.9378
Cresci-2017 std 0.0001 0.0001 0.0003 0.0001
Cresci-stock-2018 mean 0.5002 0.5073 0.7116 0.6075
Cresci-stock-2018 std 0.0002 0.0003 0.0007 0.0006
Cresci-rtbust-2019 mean 0.5118 0.4882 0.8125 0.6087
Cresci-rtbust-2019 std 0.0002 0.0001 0.0009 0.0003
Midterm-2018 mean 0.8459 0.8530 0.9864 0.9148
Midterm-2018 std 0.0000 0.0000 0.0000 0.0000
Botometer-feedback-2019 mean 0.7698 0.2500 0.1778 0.2049
Botometer-feedback-2019 std 0.0002 0.0006 0.0006 0.0006
baseline acc on Twibot-22 f1 on Twibot-22 type tags
Hayawi et al. 0.7650 0.2474 F lstm