Kudugunta

Deep neural networks for bot detection

---

• authors: Sneha Kudugunta, Emilio Ferrara

• link: https://arxiv.org/abs/1802.04289

• file structure:

└── train.py # train model on every dataset

• implement details: “Favorite count” is discarded since required information is not included in datasets.

How to reproduce:

1. train random forest model by and specify the dataset by running:

   python train.py --datasets ${dataset} > result.txt

   the final result will be saved into result.txt

Result:

random seed: 100, 200, 300, 400, 500

dataset		acc	precison	recall	f1
cresci-2015	mean	0.7533	1.0000	0.6095	0.7574
Cresci-2015	std	0.0013	0.0000	0.0021	0.0016
Twibot-20	mean	0.5959	0.8040	0.3347	0.4726
Twibot-20	std	0.0065	0.0060	0.0130	0.0135
Twibot-22	mean	0.6587	0.4431	0.6198	0.5167
Twibot-22	std	0.0000	0.0000	0.0000	0.0000
cresci-2017	mean	0.8832	0.9853	0.8588	0.9174
cresci-2017	std	0.0021	0.0019	0.0037	0.0017
cr-2019	mean	0.6294	0.6609	0.5067	0.4922
cr-2019	std	0.0081	0.0235	0.0121	0.0128
bf-2019	mean	0.7396	0.5667	0.4533	0.4961
bf-2019	std	0.0470	0.1077	0.0869	0.0820
cs-2018	mean	0.7753	0.5487	0.4754	0.5094
cs-2018	std	0.0014	0.0047	0.0060	0.0038
midterm-2018	mean	0.9109	0.9906	0.9024	0.9445
midterm-2018	std	0.0049	0.0016	0.0066	0.0032
gilani-2017	mean	0.7004	0.8544	0.3514	0.4975
gilani-2017	std	0.0105	0.0242	0.0170	0.0210

baseline	acc on Twibot-22	f1 on Twibot-22	type	tags
Kudugunta et al.	0.6587	0.5167	F	SMOTENN, random forest