Datasets:
dainis-boumber
commited on
Commit
•
64b9954
1
Parent(s):
f95872a
fixed numbers
Browse files
README.md
CHANGED
@@ -78,25 +78,32 @@ There are 7 independent domains in the dataset.
|
|
78 |
Each task is (or has been converted to) a binary classification problem where `y` is an indicator of deception.
|
79 |
|
80 |
1) **Phishing** (2020 Email phishing benchmark with manually labeled emails)
|
|
|
81 |
*- total: 15272 deceptive: 6074 non-deceptive: 9198*
|
82 |
|
83 |
2) **Fake News** (News Articles)
|
|
|
84 |
*- total: 20456 deceptive: 8832 non-deceptive: 11624*
|
85 |
|
86 |
3) **Political Statements** (Claims and statements by politicians and other entities, made from Politifact by relabeling LIAR)
|
|
|
87 |
*- total: 12497 deceptive: 8042 non-deceptive: 4455*
|
88 |
|
89 |
4) **Product Reviews** (Amazon product reviews)
|
|
|
90 |
*- total: 20971 deceptive: 10492 non-deceptive: 10479*
|
91 |
|
92 |
5) **Job Scams** (Job postings on an online board)
|
|
|
93 |
*- total: 14295 deceptive: 599 non-deceptive: 13696*
|
94 |
|
95 |
6) **SMS** (combination of SMS Spam from UCI repository and SMS Phishing datasets)
|
96 |
-
|
|
|
97 |
|
98 |
7) **Twitter Rumours** (Collection of rumours from PHEME dataset, covers multiple topics)
|
99 |
-
|
|
|
100 |
|
101 |
Each one was constructed from one or more datasets. Some tasks were not initially binary and had to be relabeled.
|
102 |
The inputs vary wildly both stylistically and syntactically, as well as in terms of the goal of deception
|
|
|
78 |
Each task is (or has been converted to) a binary classification problem where `y` is an indicator of deception.
|
79 |
|
80 |
1) **Phishing** (2020 Email phishing benchmark with manually labeled emails)
|
81 |
+
|
82 |
*- total: 15272 deceptive: 6074 non-deceptive: 9198*
|
83 |
|
84 |
2) **Fake News** (News Articles)
|
85 |
+
|
86 |
*- total: 20456 deceptive: 8832 non-deceptive: 11624*
|
87 |
|
88 |
3) **Political Statements** (Claims and statements by politicians and other entities, made from Politifact by relabeling LIAR)
|
89 |
+
|
90 |
*- total: 12497 deceptive: 8042 non-deceptive: 4455*
|
91 |
|
92 |
4) **Product Reviews** (Amazon product reviews)
|
93 |
+
|
94 |
*- total: 20971 deceptive: 10492 non-deceptive: 10479*
|
95 |
|
96 |
5) **Job Scams** (Job postings on an online board)
|
97 |
+
|
98 |
*- total: 14295 deceptive: 599 non-deceptive: 13696*
|
99 |
|
100 |
6) **SMS** (combination of SMS Spam from UCI repository and SMS Phishing datasets)
|
101 |
+
|
102 |
+
*- total: 6574 deceptive: 1274 non-deceptive: 5300*
|
103 |
|
104 |
7) **Twitter Rumours** (Collection of rumours from PHEME dataset, covers multiple topics)
|
105 |
+
|
106 |
+
*- total: 5789 deceptive: 1969 non-deceptive: 3820*
|
107 |
|
108 |
Each one was constructed from one or more datasets. Some tasks were not initially binary and had to be relabeled.
|
109 |
The inputs vary wildly both stylistically and syntactically, as well as in terms of the goal of deception
|