ayourway / Columbia-SEC-Pricing-NLP-ML

By conducting sentiment analysis on a company's SEC filings and pairing with its stock price, the project explores machine learning models that best predicts the price of a stock.
6 stars 5 forks source link

Columbia-SEC-Pricing-NLP-ML

The project aims to build a stock pricing prediction KPI from SEC filings

1.Scrap and parse the text part of financial reports.
2.Do sentiment analysis using NLP.
3.Access stock price data.
4.Build machine learning models using sentiment data and stock price.

Contributors of the project are:

- Jieyu Yao <y.jieyu@columbia.edu>,
- Yue Chang <ulysses.cy@gmail.com>,
- Qing Ma <qm2124@columbia.edu>,
- Jun Guo <jg3555@columbia.edu>.

Update Feb. 23 2018

To execute the executor.py, which scrapes SEC files using information from data/edgar.csv and converts the 10-K/10-Q documents into sentiment vectors:

python executor.py 

Thanks for the help from our professors Dr. Eugene Wu and Dr. Sambit Sahu.

The original repository is here: CYUlysses/W4121_SECNLP