Datasets:
Commit
·
9cb2890
1
Parent(s):
49fe049
update descriptions
Browse files
README.md
CHANGED
|
@@ -1,38 +1,49 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
language:
|
| 4 |
-
-
|
| 5 |
-
-
|
| 6 |
-
-
|
| 7 |
-
-
|
| 8 |
-
- tl
|
| 9 |
-
- cs
|
| 10 |
-
- da
|
| 11 |
-
- vi
|
| 12 |
-
- nl
|
| 13 |
-
- lv
|
| 14 |
-
- tr
|
| 15 |
-
- sl
|
| 16 |
- af
|
| 17 |
-
-
|
| 18 |
-
-
|
|
|
|
| 19 |
- cy
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
- sq
|
|
|
|
|
|
|
|
|
|
| 21 |
- hr
|
| 22 |
-
- lt
|
| 23 |
-
- et
|
| 24 |
- en
|
| 25 |
-
-
|
| 26 |
-
-
|
| 27 |
-
-
|
| 28 |
-
-
|
| 29 |
-
-
|
|
|
|
|
|
|
|
|
|
| 30 |
- bn
|
| 31 |
-
- pl
|
| 32 |
-
- ca
|
| 33 |
-
- sk
|
| 34 |
-
- pt
|
| 35 |
- ro
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 36 |
- multilingual
|
| 37 |
task_categories:
|
| 38 |
- text-classification
|
|
@@ -48,7 +59,9 @@ size_categories:
|
|
| 48 |
<br>
|
| 49 |
|
| 50 |
|
| 51 |
-
This is a large corpus of
|
|
|
|
|
|
|
| 52 |
|
| 53 |
<br>
|
| 54 |
|
|
@@ -59,7 +72,7 @@ This is a large corpus of 41,578 preprocessed sms messages and emails sent by hu
|
|
| 59 |
<br>
|
| 60 |
|
| 61 |
### Data composition
|
| 62 |
-
![Spam vs Non-spam (Ham)](https://i.imgur.com/
|
| 63 |
|
| 64 |
<br>
|
| 65 |
|
|
@@ -72,7 +85,7 @@ text = email.subject + "\n\n" + email.content
|
|
| 72 |
|
| 73 |
<br>
|
| 74 |
|
| 75 |
-
###
|
| 76 |
- https://huggingface.co/datasets/sms_spam
|
| 77 |
- https://github.com/MWiechmann/enron_spam_data
|
| 78 |
- https://github.com/stdlib-js/datasets-spam-assassin
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
language:
|
| 4 |
+
- no
|
| 5 |
+
- es
|
| 6 |
+
- so
|
| 7 |
+
- ca
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
- af
|
| 9 |
+
- it
|
| 10 |
+
- nl
|
| 11 |
+
- hi
|
| 12 |
- cy
|
| 13 |
+
- ar
|
| 14 |
+
- sv
|
| 15 |
+
- cs
|
| 16 |
+
- pl
|
| 17 |
+
- de
|
| 18 |
+
- lt
|
| 19 |
- sq
|
| 20 |
+
- uk
|
| 21 |
+
- tl
|
| 22 |
+
- sl
|
| 23 |
- hr
|
|
|
|
|
|
|
| 24 |
- en
|
| 25 |
+
- fi
|
| 26 |
+
- vi
|
| 27 |
+
- id
|
| 28 |
+
- da
|
| 29 |
+
- ko
|
| 30 |
+
- bg
|
| 31 |
+
- mr
|
| 32 |
+
- ja
|
| 33 |
- bn
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
- ro
|
| 35 |
+
- pt
|
| 36 |
+
- fr
|
| 37 |
+
- hu
|
| 38 |
+
- tr
|
| 39 |
+
- zh
|
| 40 |
+
- mk
|
| 41 |
+
- ur
|
| 42 |
+
- sk
|
| 43 |
+
- ne
|
| 44 |
+
- et
|
| 45 |
+
- sw
|
| 46 |
+
- ru
|
| 47 |
- multilingual
|
| 48 |
task_categories:
|
| 49 |
- text-classification
|
|
|
|
| 59 |
<br>
|
| 60 |
|
| 61 |
|
| 62 |
+
This is a large corpus of 42,619 preprocessed text messages and emails sent by humans in 42 languages. `is_spam=1` means spam and `is_spam=0` means ham.
|
| 63 |
+
|
| 64 |
+
1040 rows of balanced data, consisting of casual conversations and scam emails in ≈10 languages, were manually collected and annotated by me, with some help from ChatGPT.
|
| 65 |
|
| 66 |
<br>
|
| 67 |
|
|
|
|
| 72 |
<br>
|
| 73 |
|
| 74 |
### Data composition
|
| 75 |
+
|
| 76 |
|
| 77 |
<br>
|
| 78 |
|
|
|
|
| 85 |
|
| 86 |
<br>
|
| 87 |
|
| 88 |
+
### Other Sources
|
| 89 |
- https://huggingface.co/datasets/sms_spam
|
| 90 |
- https://github.com/MWiechmann/enron_spam_data
|
| 91 |
- https://github.com/stdlib-js/datasets-spam-assassin
|