Datasets:
dainis-boumber
commited on
Commit
•
4db0676
1
Parent(s):
cd17ca5
Update README.md
Browse files
README.md
CHANGED
@@ -60,47 +60,82 @@ configs:
|
|
60 |
|
61 |
# GDDs-2.0
|
62 |
|
63 |
-
The Generalized Deception Dataset version 2.0 is a labeled corpus containing over 95000 samples of
|
|
|
64 |
|
65 |
## Authors
|
66 |
|
67 |
ReDAS Lab, University of Houston, 2023. See https://www2.cs.uh.edu/~rmverma/ for contact information.
|
68 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
69 |
## Contents
|
70 |
|
71 |
Each task is (or has been converted to) a binary classification problem.
|
72 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
73 |
Each subdirectory/config contains the domain/individual dataset.
|
74 |
`train.jsonl`, `test.jsonl`, and `valid.jsonl` contain train, test, and validation sets, respectively.
|
75 |
|
76 |
The splits are train=80%, test=10%, valid=10%
|
77 |
The sampling process was random with seed=42, and stratified with respect to `y` (label) for each domain.
|
78 |
|
|
|
|
|
79 |
Each `jsonl` file has two fields (columns): `text` and `label`
|
80 |
|
81 |
`label` answers the question whether text is deceptive: `1` means yes, it is deceptive, `0` means no, the text is not deceptive (it is truthful).
|
82 |
|
83 |
`text` is guaranteed to be valid unicode, less than 1 million characters, and contains no empty entries or non-values.
|
84 |
|
85 |
-
|
86 |
|
87 |
-
|
|
|
|
|
88 |
|
89 |
-
## List of Domains/Datasets
|
90 |
-
|
91 |
-
1) Phishing (Email)
|
92 |
-
|
93 |
-
2) Fake News (News Articles)
|
94 |
-
|
95 |
-
3) Political Statements (Claims and statements by politicians and other entities, made from LIAR)
|
96 |
-
|
97 |
-
4) Product Reviews (Amazon product reviews)
|
98 |
-
|
99 |
-
5) Job Scams (Job postings)
|
100 |
-
|
101 |
-
6) SMS (Phishing attacks via sms) (combination of SMS Spam from UCI repository and SMS Phishing datasets)
|
102 |
-
|
103 |
-
7) Twitter Rumours (Collection of rumours from PHEME dataset, covers multiple topics)
|
104 |
|
105 |
## Changes and Additions
|
106 |
|
@@ -142,17 +177,19 @@ sms total: 6574 deceptive: 1274 non-deceptive: 5300
|
|
142 |
|
143 |
twitter_rumours total: 5789 deceptive: 1969 non-deceptive: 3820
|
144 |
|
|
|
145 |
|
146 |
-
|
|
|
147 |
|
148 |
-
|
149 |
-
|
150 |
-
## Citing
|
151 |
|
152 |
If you found this dataset useful in your research, please consider citing it as:
|
153 |
|
154 |
TODO: ADD our paper reference
|
155 |
|
|
|
|
|
156 |
Original GDD paper:
|
157 |
|
158 |
@inproceedings{10.1145/3508398.3519358,
|
@@ -174,6 +211,8 @@ series = {CODASPY '22}
|
|
174 |
}
|
175 |
|
176 |
|
|
|
|
|
177 |
## APPENDIX: Dataset and Domain Details
|
178 |
|
179 |
This section describes each domain/dataset in greater detail.
|
|
|
60 |
|
61 |
# GDDs-2.0
|
62 |
|
63 |
+
The Generalized Deception Dataset version 2.0 is a labeled corpus containing over 95000 samples of
|
64 |
+
deceptive and truthful texts from a number of independent domains and tasks.
|
65 |
|
66 |
## Authors
|
67 |
|
68 |
ReDAS Lab, University of Houston, 2023. See https://www2.cs.uh.edu/~rmverma/ for contact information.
|
69 |
|
70 |
+
## Domains/Sub-Tasks
|
71 |
+
|
72 |
+
There are 7 independent domains in the dataset. Each one was constructed from one or more datasets.
|
73 |
+
Some tasks were not initially binary and had to be relabeled. The inputs vary wildly both stylistically
|
74 |
+
and syntactically, as well as in terms of the goal of deception (or absence of thereof) being performed in the context of each
|
75 |
+
dataset. The two uniting factors are: all seven datasets contain some fraction of texts that are meant
|
76 |
+
to deceive the person reading them one way or another.
|
77 |
+
|
78 |
+
1) Phishing (Email)
|
79 |
+
|
80 |
+
2) Fake News (News Articles)
|
81 |
+
|
82 |
+
3) Political Statements (Claims and statements by politicians and other entities, made from LIAR)
|
83 |
+
|
84 |
+
4) Product Reviews (Amazon product reviews)
|
85 |
+
|
86 |
+
5) Job Scams (Job postings)
|
87 |
+
|
88 |
+
6) SMS (Phishing attacks via sms) (combination of SMS Spam from UCI repository and SMS Phishing datasets)
|
89 |
+
|
90 |
+
7) Twitter Rumours (Collection of rumours from PHEME dataset, covers multiple topics)
|
91 |
+
|
92 |
## Contents
|
93 |
|
94 |
Each task is (or has been converted to) a binary classification problem.
|
95 |
|
96 |
+
### Structure
|
97 |
+
|
98 |
+
The directory layout of gdds is like so:
|
99 |
+
|
100 |
+
``
|
101 |
+
gdds
|
102 |
+
fake_news/
|
103 |
+
train.jsonl
|
104 |
+
test.jsonl
|
105 |
+
validation.jsonl
|
106 |
+
README.md
|
107 |
+
...
|
108 |
+
...
|
109 |
+
...
|
110 |
+
sms/
|
111 |
+
train.jsonl
|
112 |
+
test.jsonl
|
113 |
+
validation.jsonl
|
114 |
+
README.md
|
115 |
+
README.md
|
116 |
+
LICENSE.txt
|
117 |
+
``
|
118 |
+
|
119 |
Each subdirectory/config contains the domain/individual dataset.
|
120 |
`train.jsonl`, `test.jsonl`, and `valid.jsonl` contain train, test, and validation sets, respectively.
|
121 |
|
122 |
The splits are train=80%, test=10%, valid=10%
|
123 |
The sampling process was random with seed=42, and stratified with respect to `y` (label) for each domain.
|
124 |
|
125 |
+
### Fields
|
126 |
+
|
127 |
Each `jsonl` file has two fields (columns): `text` and `label`
|
128 |
|
129 |
`label` answers the question whether text is deceptive: `1` means yes, it is deceptive, `0` means no, the text is not deceptive (it is truthful).
|
130 |
|
131 |
`text` is guaranteed to be valid unicode, less than 1 million characters, and contains no empty entries or non-values.
|
132 |
|
133 |
+
### Documentation
|
134 |
|
135 |
+
Primary documentation is this README file. Each dataset's directory contains a `README.md` file with additional details.
|
136 |
+
The contents of these files are also included at the end of this document in the Appendix.
|
137 |
+
LICENSE.txt contains the MIT license this dataset is distributed under.
|
138 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
139 |
|
140 |
## Changes and Additions
|
141 |
|
|
|
177 |
|
178 |
twitter_rumours total: 5789 deceptive: 1969 non-deceptive: 3820
|
179 |
|
180 |
+
## LICENSE
|
181 |
|
182 |
+
This dataset is published under the MIT license and can be used and modified by anyone free of charge.
|
183 |
+
See LICENSE.txt file for details.
|
184 |
|
185 |
+
## CITING
|
|
|
|
|
186 |
|
187 |
If you found this dataset useful in your research, please consider citing it as:
|
188 |
|
189 |
TODO: ADD our paper reference
|
190 |
|
191 |
+
## REFERENCES
|
192 |
+
|
193 |
Original GDD paper:
|
194 |
|
195 |
@inproceedings{10.1145/3508398.3519358,
|
|
|
211 |
}
|
212 |
|
213 |
|
214 |
+
|
215 |
+
|
216 |
## APPENDIX: Dataset and Domain Details
|
217 |
|
218 |
This section describes each domain/dataset in greater detail.
|