Name of QuantLet : TXTmedia_sentiment_stock_market_prediction
Published in : TXT
Description : Prediction of Apple Inc. stock return using BERT-based sentiment index built from NASDAQ News data.
The sentiment classification algorithms (namely BERT and Support Vector Machine) are trained using Malo et al. (2014)’s « Sentences_66Agree.txt ». The lexicon-based index is built using Loughran and MacDonald (2011)'s dictionary.
We use LSTM (Hochreiter, S. and J. Schmidhuber(1997)) for prediction and compare the performance depending on the inputs.
The controls are the one-month USD LIBOR interest rate, the economic risk premium defined by the difference between the
three-month and the one-month interest rates, the USD currency fluctuation using the US dollar index (USDX) as a proxy
and the Fama-French 5 factors.
The data for the AAPL stock value and the USD index can be found on Yahoo Finance
The data for the US LIBOR interest rates and the risk premium can be found on
http://iborate.com/usd-libor/.
The data for the Fama-French 5 factors can be found on
http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/Data_Library/f-f_5_factors_2x3.html
The NASDAQ News dataset was built by Junjie Hu from the Ladislaus von Bortkiewicz Chair of Statistics of the Humboldt-Universität zu Berlin. This database is available for research purposes at RDC, CRC 649, Humboldt-Universität zu Berlin
The tutorial used for LSTM prediction can be found here:
https://machinelearningmastery.com/multivariate-time-series-forecasting-lstms-keras/
Keywords : stock market prediction, media sentiment, transfer learning, BERT, support vector machine
See also : TXTfpbsupervisedBERT for BERT sentiment classification performance evaluation
Author : Manuel Tonneau
Submitted : Thu, Sep 13 2019 by Manuel Tonneau
manueltonneau/TXTmedia_sentiment_stock_market_prediction
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
About
No description, website, or topics provided.
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published