Fake Detect: A Deep Learning Ensemble Model for Fake News Detection

Table 1

Description of the LIAR dataset.

No.	Feature name	Datatype	Missing values	Mean (μ) ± Std (σ)	Range	No. of categories
1	ID of the statement	Object	—	—
2	Label	Object	—	—		2
3	Statement	Object	—	—		4007
4	Subject(s)	Object	—	—		1823
5	Speaker	Object	—	—		9
6	Speaker’s job title	Object	1184	—		656
7	State info	Object	926	—
8	Party affiliation	Object	—	—		4
9	Barely true counts	Int (64)	—	11.59 ± 18.98	0–70
10	False counts	Int (64)	—	13.36 ± 24.14	0–114
11	Half true counts	Int (64)	—	17.19 ± 35.85	0–160
12	Mostly true counts	Int (64)	—	16.50 ± 36.17	0–163
13	Pants on fire counts	Int (64)	—	6.25 ± 16.18	0–70
14	The context (venue/location of the speech or statement)	Object	52