Finance AI: 7 of the Best Financial AI Tools For Everyday Investors & Developers
The goal here is to explore artificial intelligence financial tools that are publicly available for anyone to use
Check out the following financial AI tools on this page
- FinBERT
- FinBERT Tone
- FinBERT Tone Fine-Tuned FinTwitter Classification
- FinBERT Tone Fine-Tuned Finance Topic Classification
- Financial BERT
- FinBERT FinnSentiment
- Financial BERT-QA
FinBERT
FinBERT is a pre-trained NLP model that analyzes the sentiment of financial text
You feed FinBERT financial text and it responds with sentiment scores for positive, negative, and neutral
See how FinBERT works
Notes on FinBERT
- Built upon BERT, FinBERT is further pre-trained using 46,143 Reuters financial documents with over 29M words and nearly 400K sentences
- FinBERT is fine-tuned for financial sentiment classification using a large financial corpus
- The financial corpus is the Financial Phrase Bank published in Good Debt or Bad Debt: Detecting Semantic Orientations in Economic Texts by Malo (2014)
- In contrast, BERT was pre-trained on a general set of text from the English Wikipedia and BooksCorpus
- FinBERT beat other existing algorithms and models in the 2019 era but nonetheless left room for improvement
- One area FinBERT excelled at is correctly predicting positive and negative sentiments when others predicted neutral
- FinBERT's results show that smaller datasets can effectively further pre-train and fine-tune existing large-language models for domains containing unique vocabulary
- FinBERT beats algorithms because FinBERT uses contextual information, whereas algorithms focus on the existence of words and phrases
- FinBERT beats pre-existing models because FinBERT is further pre-trained and fine-tuned specifically on financial texts, whereas pre-existing models are more general
Links for FinBERT With Results, Examples, and Data
FinBERT Tone
FinBERT Tone is a pre-trained NLP model that analyzes the sentiment of financial text
You feed FinBERT Tone financial text and it responds with sentiment scores for positive, negative, and neutral
See how FinBERT Tone works
Notes on FinBERT Tone
- FinBERT Tone is a highly specific model fine-tuned on a small amount of data
- Built upon FinBERT, FinBERT Tone is fine-tuned on 10,000 analyst reports that are manually marked by humans as positive, negative, or neutral
- FinBERT Tone reports superior performance over FinBERT for financial sentiment analysis
- FinBERT Tone shows that smaller datasets can effectively fine-tune existing large-language models for domains like finance containing words infrequently used elsewhere
- FinBERT Tone beats pre-existing models and algorithms at financial sentiment analysis because it is so specifically fine-tuned for financial text sentiment analysis
- Like FinBERT, FinBERT Tone excels at identifying positive or negative sentiments that others mislabel as neutral
Links for FinBERT Tone With Results, Examples, and Data
FinBERT Tone Fine-Tuned FinTwitter Classification
FinBERT Tone Fine-Tuned FinTwitter Classification is a pre-trained NLP model that analyzes the sentiment of Tweets
You feed FinBERT Tone Fine-Tuned FinTwitter Classification a Tweet and it responds with sentiment scores for bullish, neutral, and bearish
See how FinBERT Tone Fine-Tuned FinTwitter Classification works
The threat of a natural gas supply crunch in Europe is growing after a key price spread reversed, signaling that more cargoes will be diverted to Asia https://t.co/f4HLcxvUT7
— Bloomberg Markets (@markets) May 26, 2023
Get Predictions
with FinBERT Tone Fine-Tuned FinTwitter Classification
The pound’s rebound looks ripe for a reversal, according to BlueBay Asset Management, as stubborn price pressures reignite talk of stagflation among analysts https://t.co/FFMUCozY01
— Bloomberg Markets (@markets) May 26, 2023
Get Predictions
with FinBERT Tone Fine-Tuned FinTwitter Classification
TotalEnergies suffers investor revolt over climate goals https://t.co/s5Qp9RXjyE
— Financial Times (@FT) May 26, 2023
Get Predictions
with FinBERT Tone Fine-Tuned FinTwitter Classification
Notes on FinBERT Tone Fine-Tuned FinTwitter Classification
- Built upon FinBERT Tone, FinBERT Tone Fine-Tuned FinTwitter Classification is fine-tuned on the Twitter Financial News Dataset
- Review the other sections to learn more about FinBERT Tone, FinBERT, and BERT
Links for FinBERT Tone Fine-Tuned FinTwitter Classification With Results, Examples, and Data
FinBERT Tone Fine-Tuned Finance Topic Classification
FinBERT Tone Fine-Tuned Finance Topic Classification is a pre-trained NLP model that classifies financial Tweets
You feed FinBERT Tone Fine-Tuned Finance Topic Classification a Tweet and it responds with scores for 20 different financial topics
See how FinBERT Tone Fine-Tuned Finance Topic Classification works
The threat of a natural gas supply crunch in Europe is growing after a key price spread reversed, signaling that more cargoes will be diverted to Asia https://t.co/f4HLcxvUT7
— Bloomberg Markets (@markets) May 26, 2023
Get Predictions
with FinBERT Tone Fine-Tuned Finance Topic Classification
The pound’s rebound looks ripe for a reversal, according to BlueBay Asset Management, as stubborn price pressures reignite talk of stagflation among analysts https://t.co/FFMUCozY01
— Bloomberg Markets (@markets) May 26, 2023
Get Predictions
with FinBERT Tone Fine-Tuned Finance Topic Classification
TotalEnergies suffers investor revolt over climate goals https://t.co/s5Qp9RXjyE
— Financial Times (@FT) May 26, 2023
Get Predictions
with FinBERT Tone Fine-Tuned Finance Topic Classification
Notes on FinBERT Tone Fine-Tuned Finance Topic Classification
- Use FinBERT Tone Fine-Tuned Finance Topic Classification to create automations or help you quickly understand the topics of financial Tweets
- Built upon FinBERT Tone, FinBERT Tone Fine-Tuned Finance Topic Classification is fine-tuned using the Twitter Financial News Topic dataset
Links for FinBERT Tone Fine-Tuned Finance Topic Classification With Results, Examples, and Data
Financial BERT
Financial BERT is a pre-trained NLP model that classifies financial text
You feed Financial BERT financial text and it responds with scores for keywords telling you what the text is about
See how Financial BERT works
Try it out
Financial BERT
Notes on Financial BERT
- Financial BERT has the same structure as BERT but instead of being pre-trained on English Wikipedia and BooksCorpus like BERT, Financial BERT is pre-trained for financial text classification using a large financial corpus
- The financial corpus for pre-training Financial BERT consists of 1.8M Reuters News articles (from 2008 to 2010), 400,000 Bloomberg News articles (from 2006 to 2013), 192,000 Corporate Reports (10-K & 10-Q), and 42,156 Earning Calls
- Financial BERT reports better scores than financial NLP models built on pre-trained BERT models, like FinBERT
Links for Financial BERT With Results, Examples, and Data
FinBERT FinnSentiment
FinBERT FinnSentiment is a pre-trained NLP model that analyzes the sentiment of financial text
You feed FinBERT FinnSentiment financial text and it responds with sentiment scores for positive, negative, and neutral
See how FinBERT FinnSentiment works
Notes on FinBERT FinnSentiment
- Built upon FinBERT, FinBERT FinnSentiment is fine-tuned on the FinnSentiment dataset
- Review the other sections to learn more about FinBERT and BERT
Links for FinBERT FinnSentiment With Results, Examples, and Data
Financial BERT-QA
FinBERT-QA is a question-and-answering system for asking financial questions and getting an opinionated financial answer in response
You feed FinBERT-QA a financial question and it responds with a financial answer
See how Financial BERT-QA works
Try it out
FinBERT-QA
Notes on Financial BERT-QA
- Use FinBERT-QA as part of a chat bot or in automations
- FinBERT-QA is built by transferring and fine-tuning a pre-trained BERT model for general question and answering, and then further fine-tuning the resulting model using the FiQA dataset to orient it toward finance