- תיאור:
Sentiment140 מאפשר לך לגלות את הרגש של מותג, מוצר או נושא בטוויטר.
הנתונים הם קובץ CSV עם הסברים של סמלי הבעה. פורמט קובץ הנתונים כולל 6 שדות:
- הקוטביות של הציוץ (0 = שלילי, 2 = ניטרלי, 4 = חיובי)
- מזהה הציוץ (2087)
- תאריך הציוץ (שבת 16 במאי 23:58:44 UTC 2009)
- השאילתה (lyx). אם אין שאילתה, אז ערך זה הוא NO_QUERY.
- המשתמש שצייץ (robotickilldozr)
- הטקסט של הציוץ (Lyx מגניב)
לקבלת מידע נוסף, עיינו סיווג סנטימנט נייר טוויטר עם פיקוח רחוק בלב https://cs.stanford.edu/people/alecmgo/papers/TwitterDistantSupervision09.pdf
דף הבית: http://help.sentiment140.com/home
קוד מקור:
tfds.text.sentiment140.Sentiment140
גרסאות:
-
1.0.0
(ברירת המחדל): הערות השחרור לא.
-
גודל ההורדה:
77.59 MiB
מערך נתונים גודל:
305.13 MiB
Auto-במטמון ( תיעוד ): אין
פיצולים:
לְפַצֵל | דוגמאות |
---|---|
'test' | 498 |
'train' | 1,600,000 |
- מאפיינים:
FeaturesDict({
'date': Text(shape=(), dtype=tf.string),
'polarity': tf.int32,
'query': Text(shape=(), dtype=tf.string),
'text': Text(shape=(), dtype=tf.string),
'user': Text(shape=(), dtype=tf.string),
})
מפתחות השגחה (ראה
as_supervised
doc ):('text', 'polarity')
איור ( tfds.show_examples ): לא נתמך.
דוגמאות ( tfds.as_dataframe ):
- ציטוט:
@ONLINE {Sentiment140,
author = "Go, Alec and Bhayani, Richa and Huang, Lei",
title = "Twitter Sentiment Classification using Distant Supervision",
year = "2009",
url = "http://help.sentiment140.com/home"
}