Miller

Twitter spammer detection using data stream clustering

---

• authors: Zachary Miller, Brian Dickinson, William Deitrick, Wei Hu, Alex Hai Wang

• link: https://dl.acm.org/doi/10.1016/j.ins.2013.11.016

• file structure:

├── preprocess.py       # generate required features
└── stream_cluster.py   # train model on given dataset

• implement details: “Re-tweet count” feature is discarded since required information is not included in datasets. “Link count”, “Reply/mention count”, and “Hashtag count” are available only in “cresci-2015”, “cresci-2017”, “Twibot-20”, and “Twibot-22” datasets.

How to reproduce:

1. specify the dataset by running dataset=Twibot-22 (Twibot-22 for example) ;

2. generate required features from raw dataset by running:

   python preprocess.py --dataset ${dataset}

3. train clustering model by running:

   python stream_cluster.py --dataset ${dataset}

Result:

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

dataset		acc	precison	recall	f1
Cresci-2015	mean	0.7551	0.7207	1.0000	0.8377
Cresci-2015	std	0.0000	0.0000	0.0000	0.0000
Cresci-2017	mean	0.7713	0.7721	0.9911	0.8680
Cresci-2017	std	0.0017	0.0018	0.0011	0.0007
Twibot-20	mean	0.6450	0.6071	0.9744	0.7481
Twibot-20	std	0.0035	0.0020	0.0047	0.0026
Twibot-22	mean	0.3037	0.2946	0.9789	0.4529
Twibot-22	std	0.0001	0.0000	0.0001	0.0000
botometer-feedback-2019	mean	0.7736	0.0000	0.0000	0.0000
botometer-feedback-2019	std	0.0000	0.0000	0.0000	0.0000
cresci-rtbust-2019	mean	0.5441	0.5217	0.3750	0.4364
cresci-rtbust-2019	std	0.0000	0.0000	0.0000	0.0000
cresci-stock-2018	mean	0.5253	0.5478	0.5889	0.5676
cresci-stock-2018	std	0.0000	0.0000	0.0000	0.0000
midterm-2018	mean	0.8372	0.8385	0.9981	0.9114
midterm-2018	std	0.0000	0.0000	0.0000	0.0000
gilani-2017	mean	0.5104	0.4889	0.7719	0.5986
gilani-2017	std	0.0000	0.0000	0.0000	0.0000

baseline	acc on Twibot-22	f1 on Twibot-22	type	tags
Miller et al.	0.3037	0.4529	F T	k means