A library for market sentiment analysis of financial social media posts
Project description
pyFin-Sentiment
A library for sentiment analysis of financial social media posts
Sentiment Analysis of Financial Social Media Posts
This library can help you extract market sentiment from short social media posts. Trained on data from Twitter, it can classify sentimetn into three classes: Bullish, Bearish, Neutral/No Sentiment. Note that we need to differentiate between market sentiment and general sentiment. Consider this example:
💬 Nice, already made loads of money this morning and now im shorting $AAPL, let's goooo!
While the general sentiment in the text is positve, the market sentiment is negative as the author is shorting a stock. Therefore, ...
- If you are looking for a generic sentiment model that works well on social media content, take a look at VADER or TwitterRoBERTa
- If you are looking for a sentiment analysis models that excels on new headlines sentiment analysis, check out FinBERT
- Otherwise, stay here 🙃
Installation
It's as easy as...
pip install pyfin-sentiment
Documentation
📚 The documentation lives on pyfin-sentiment.readthedocs.io
Example
from pyfin_sentiment.model import SentimentModel
# the model only needs to be downloaded once
SentimentModel.download("small")
model = SentimentModel("small")
model.predict(["Long $TSLA!!", "Selling my $AAPL position"])
# array(['1', '3'], dtype=object)
We use the following conventions for mapping sentiment classes:
Class Name | Meaning |
---|---|
1 | Positive, Bullish |
2 | Neutral, Uncertain |
3 | Negative, Bearish |
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for pyfin_sentiment-0.1.1-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 1d9e7687b2cae00d614add5a67bc44195be32420bee1f516b79be90487ea11c9 |
|
MD5 | fd257deba08962e872b2e23fea340265 |
|
BLAKE2b-256 | b6be05c3f6c4de46c0454088d9db8b6c1916fcb7ec56167d311bf4c0ab7efe85 |