Hey guys, Hey girls,
As I want to start coding a new Trading AI in this year (first based on Python and later maybe in C++) I stumbled over the following question:
Today I would like to make a pro/contra list with you in the area of Deep- vs Machine Learning. The difference should be clear to most of you if not here is a nice excerpt from Hackernoon:
Machine Learning for dummies:
A subset of artificial intelligence involved with the creation of algorithms that can modify itself without human intervention to produce desired output- by feeding itself through structured data.
Deep Learning for dummies:
A subset of machine learning where algorithms are created and function similar to those in machine learning, but there are numerous layers of these algorithms- each providing a different interpretation to the data it feeds on. Such a network of algorithms is called artificial neural networks, being named so as their functioning is an inspiration, or you may say; an attempt at imitating the function of the human neural networks present in the brain.
Up to now I was convinced that this future project will be based on Tensorflow, Keras, etc. However, the following came to my mind afterward.
Most of you will probably have heard of Pluribus already. Dr. Sandholm and Mr. Brown (as work for his PhD) were the first to program an AI that won against 6 poker world champions in No-Limit-Texas Holdem Poker. This seemed impossible because poker is a game of imperfect information. If you haven't read/seen their work until now, I'll link this article below to the corresponding sources.
From the work, it is clear that they wrote the whole thing in C++ and WITHOUT the use of any deep learning library but exclusively on the basis of machine learning. This was also confirmed to me by both of them via email. So it was possible to bring the AI in under 24H to a level that could beat 6 Poker World Champions without any problems.
The stock and crypto market is nothing else. A game of imperfect information. The prices of a crypto coin or stock are influenced by an incredible number of factors. This includes of course prices from the past but also current media (as currently seen with covid-19) and data from the economy.
We want to grab the Data out of the "Big Players" like CoinCap API, CryptoAPI.io for all kinds of historical and new charts prices etc. The same we will do with the Yahoo Finance API to grab data out of the stock market. Depending on the size of this project and how it will develop we want to implement also some kind of NLP to grab the most out of Economy Data like dailyfx news to predict some in/decreases for some stocks but this is a future feature. So basically the main question is what would be better for our system to be used. CNNs or machine learning.
All this leads me to the conclusion that I am not sure what the better option for a trading bot would be. I know that the training of AI-based on deep learning would take much longer than based on machine learning but is it safe to say that the results are really better?
What do you think are the pros and cons foagainst Deep learning vs. Machine Learning?
I am EXTREMELY curious about your answers and look forward to an exchange with you!