Stock Prediction [0]: Make Some Moola!
Want to bank some serious bucks? Combine the Stock Market and Artificial Intelligence, and you’ve got your answer! Use deep learning for stock market prediction, and you can get some pretty stellar results!
AI is a complicated subject, comprising of seemingly-random code and an even more confusing jargon of a literature base. Fortunately, you’ve come to the right place. This blog – styled series will teach you the basics of Deep Learning, without having to learn all the complicated mathematical functions and formulas that complicate most lectures and discourage many students.
Why is this different from all the other literature targeted towards beginners in the quickly advancing field of AI? It’s simple – this whole series is a basic documentation of my own progress as a beginner at Deep Learning. At the time of the writing of the code referenced in this series, I barely knew Python, nor the endless libraries and random commands needed to make an intelligent robot. Neither do you.
What you should know are the basics of how Deep Learning works, which you can find on my short post “What is Deep Learning/AI?”. It should be a quick 3 minute read!
Several studies have shown the effect of the news on the stock market. This project was actually based on a paper done by Kalyani Joshi and 2 professors at the KJ Somaiya College of Engineering. If you are interested in making money and learning AI as you do it, this tutorial should be especially intriguing for you! This series will be focused on creating a basic model to analyze recent financial news and tell you the prediction for any ticker. We can then further this into a tool to tell you the “Hot Stocks” of the day.
When finding the best way to go about predicting Stocks, the literature base is quite thin – all these insider secrets are kept within the brains of a select few analysts on Wall Street, where hedge funds are banking big bucks because of their extremely fast News and Data-analyzing machines. We won’t be building something as complex as that today, but the goal of this project is to build a bigger system based on the basic model we build in this series.
The series will be split into 10 parts:
- Set Up, Dependencies, and Road Map
- Article Data Collection and Interpretation
- Article Analysis (1/2): Dictionaries
- Article Analysis (2/2): The Sentiment Analysis
- Stock Trend Analysis
- Generating a Dataset
- Creating a Model
- Usage Functions
- Next Steps
- Updates
Want to Follow This Series?
How is deep learning better (or worse) than what traditional hedge fund algorithms do?
Current Hedge Fund algorithms follow a set list of defined instructions for ordering a trade – this helps investors with speed. Our algorithm helps investors with research into a stock, without having to define instructions – Deep Learning allows the algorithm itself to learn what to do and when to do it.
“Traditional Hedge Fund Algorithms:”
http://www.investopedia.com/articles/active-trading/101014/basics-algorithmic-trading-concepts-and-examples.asp
Hi, thank you for the article! I’ve started the process of building a stock prediction AI and I appreciate this series of articles. I notice that not all the parts are linked, though…
I have also found some good resources at Data Hunters here. Have you encountered them before?
Hi Luisa — No problem! You can use the bottom navbar to find other posts in the series!