Load Dictionaries#
📚 Available Dictionaries#
sentibank
offers a comprehensive collection of sentiment dictionaries, including 15 original dictionaries and 43 preprocessed versions. To access the original dictionaries, you can use the predefined lexicon identifier following the {NAME}_{VERSION}
convention.
For the preprocessed dictionaries, there are two naming conventions:
{NAME}_{VERSION}
: This convention indicates that only compulsory processing has been applied to the base lexicon.{NAME}_{VERSION}_{refined}
: This structure specifies additional transformations or discretionary refinements that have been applied to the base lexicon.
For example, NoVAD_v2013_boosted
applies arousal-based adjustments to intensify extreme valence values and dampen neutral ones, providing a richness-preserving single score.
To view the available predefined lexicon identifiers and their corresponding dictionaries, please click and open the List of Available Dictionaries below.
List of Available Dictionaries
Sentiment Dictionary |
Affiliated Institution |
Description |
Genre |
Domain |
Predefined Identifiers (preprocessed) |
---|---|---|---|---|---|
AFINN |
DTU Informatics |
General purpose lexicon with sentiment ratings for common emotion words. |
Social Media |
General |
|
Aigents+ |
Autonio Foundation |
Lexicon optimised for social media posts related to cryptocurrencies. |
Social Media |
Cryptocurrency |
|
ANEW |
NIMH Center for Emotion and Attention |
Provides normative emotional ratings across pleasure, arousal, and dominance dimensions. |
General |
Psychology |
|
Dictionary of Affect in Language (DAL) |
Laurentian University |
Lexicon designed to quantify pleasantness, activation, and imagery dimensions across diverse everyday English words. |
General |
General |
|
Discrete Emotions Dictionary (DED) |
Gallup |
Lexicon focused on precisely distinguishing four key discrete emotions in political communication |
News |
Political Science |
|
General Inquirer |
Harvard University |
Lexicon capturing broad psycholinguistic dimensions across semantics, values and motivations. |
General |
Psychology, Political Science |
|
Henry |
University of Miami |
Leixcon designed for analysing tone in earnings press releases. |
Corporate Communication (Earnings Press Releases) |
Finance |
|
MASTER |
University of Notre Dame |
Financial lexicons covering expressions common in business writing. |
Regulatory Filings (10-K) |
Finance |
|
Norms of Valence, Arousal and Dominance (NoVAD) |
McMaster University |
A lexicon of 14,000 common English lemmas across valence, arousal, and dominance dimensions. |
General |
Psychology |
|
OpinionLexicon |
University of Illinois Chicago |
Opinion words tailored for sentiment analysis of product reviews. |
Product Reviews |
Consumer Products |
|
SenticNet |
Sentic Research Group |
Conceptual lexicon providing multidimensional sentiment analysis for commonsense concepts and expressions. |
General |
General |
|
SentiWordNet |
Institute of Information Science and Technologies |
Lexicon associating WordNet synsets with positive, negative, and objective scores. |
General |
General |
|
VADER |
Georgia Institute of Technology |
General purpose lexicon optimised for social media and microblogs. |
Social Media |
General |
|
WordNet-Affect |
Institute for Scientific and Technological Research |
Hierarchically organised affective labels providing a granular emotional dimension. |
General |
Psychology |
|
📖 Load Preprocessed Dictionaries#
The sentibank.archive
module provides access to 15 original and 43 preprocessed sentiment dictionaries. To load a preprocessed dictionary in dict
format:
from sentibank import archive
load = archive.load()
vader = load.dict("VADER_v2014")
{'$:': -1.5,
'%)': -0.4,
'%-)': -1.5,
'&-:': -0.4,
'&:': -0.7,
"( '}{' )": 1.6,
'(%': -0.9,
...}
📘 Load Original Dictionaries#
To load the original (unprocessed) dictionary as a pd.DataFrame
:
from sentibank import archive
load = archive.load()
afinn = load.origin("AFINN_v2015")
lexicon | score | |
---|---|---|
0 | abandon | -2 |
1 | abandoned | -2 |
2 | abandons | -2 |
3 | abducted | -2 |
4 | abduction | -2 |
... | ... | ... |
3377 | yucky | -2 |
3378 | yummy | 3 |
3379 | zealot | -2 |
3380 | zealots | -2 |
3381 | zealous | 2 |
3382 rows × 2 columns