dainis-boumber commited on
Commit
64b9954
1 Parent(s): f95872a

fixed numbers

Browse files
Files changed (1) hide show
  1. README.md +9 -2
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
- *- total: 6574 deceptive: 1274 non-deceptive: 5300*
 
97
 
98
  7) **Twitter Rumours** (Collection of rumours from PHEME dataset, covers multiple topics)
99
- *- total: 5789 deceptive: 1969 non-deceptive: 3820*
 
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