-
No$tradamu$
-
Goose-Bayes Theorem
-
the grind is real
Inspiration
Fundamental analysis is being over taken by technical analysis and quant modelling. The markets used to provide a platform for individual investors to be able to profit from making investments without being at a significant disadvantage when compared to institutional investors due to the symmetrical availability of information. Now, with technical analysis the institutions have a far greater availability of information. Even though services like IBM's Bluemix and Watson provide the data to individuals, they are unable to use it to get meaningful information. Goosefield Capital's Nostradamus allows everyone to access high level quant models and make wise investing decisions.
What it does
Uses historical financial data and sentiment analysis of news outlets to predict future equity price trends.
How we built
Nostradamus was built using python for sentiment analysis and wrapper code along with various APIs including IBM-Watson. We also used a modified version of Bayes Theorem (theorized inhouse) coded in C++ to use historical prices to forecast trends.
Challenges we ran into
Did not have experience with IBM APIs or NLP. Learned both over the course of the hackathon and had to pivot midway after realizing a better implementation.
Accomplishments that we're proud of & What we learned
Learned how to use IBM-Watson and Bluemix APIs and developed an understanding of NLP and sentiment analysis. Built a project we are both very interested in and believe it has great upside potential and future use.
What's next for Nostradamus
Implement various ML algorithms to improve efficiency. Make open-source to allow the community to help improve this project and make quant analysis available to everyone.
Log in or sign up for Devpost to join the conversation.