The Blog

Financial success using AI and Time Travel

Posted on January 18, 2018 by

Wait let me explain. I can explain every part of this click-bait title, it will make sense I promise. So, A great philosopher named Homer Simpsons once said:

"Trying is the first step towards failure"
Trying is the first step towards failure

And I agree, however Failure is the first step towards success. Learning from past mistakes is a crucial step to perfecting yourself, your skills, or your product.

Predicting Stock Movements

Let’s say you have a theory you want to test that you believe can predict if a stock is going to crash or explode. Putting your theory into rigorous tests require a lot of stocks actually crashing or exploding and that doesn’t happen every day. It will take a lot of time and effort to put your theory to the test.

Time Travel

Imagine you could go back years in time and access data and signals for countless stocks you already know that would crash or explode. You could test your theory, or countless theories against this data, since you know when an event is going to happen, and you know what happened before a specific event.

Machine Learning

The fact that you think you have a method or theory that can predict a stock movement is cute and naive, however the chances you are right are pretty slim. Why don’t you just give this job to someone who could test millions of theories a second and perfect itself in the process. Given a large set of training data (permutation of signals extracted from past events) and a known outcome (known stock movement), you can train a mathematical model that will have the best chances in predicting the outcome of a new, never before seen, data-set (the present).

Financial Success

The above notions aren’t wishful thinking, they are the exact methods utilized by many companies who are leveraging Archive to extract content from historical news articles, blog posts and past online discussions to train their AI models against known stock movements. By knowing the past (over 3 years worth of data), they are able to test and train their models and utilize the outcome to the present in order to predict the future.

Told you so

You see, the title does make sense now. I kept my promise, now it’s your turn to share this article so others could learn what you’ve just learned!

This entry was posted in Big Data, Data Extraction, Machine Learning, Technology. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *