Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
The problem with a sub forum is there just isn't that much to talk about.
Advances in Financial Machine Learning contains python code in the book but none if it is really doing machine learning.
It is taken as a given you can use a supervised regression algorithm and it is so trivial De Prado doesn't even bother going over it.
The interesting parts are all econometric based and the next book after Advances in Financial Machine Learning for me is not Deep Learning in Python but Unit Root Tests in Time Series Volume 1 by Kerry Patterson because my econometric skills are so bad.
Can you help answer these questions from other members on NexusFi?
I only trade strategies automatically generated. Both totally randome and thru Genetic Evo which is basically randome generated strategies that have been merged or migrated together. Using pure bar patterns have not worked well but using standard indicators and constants works well. Are they curvefitted, some are but they are easially spottet. Do they have large drawdowns, depends on what type of strategy it is . Do they have stagnation , Yes ofcourse. Can you know when they have a drawdown or extended stagnation.. nah no clue. But if you run enought tests You can pretty much guarantee that atleast 50 % of the strategies that are launched in a portfolio will be profitable rolling 12 months and this is enought to make some money of it.
Goldman Sachs is opening up to outsiders as the 150-year-old investment bank tries to become the Google of finance.
The bank is taking applications for a new one-year program that pays computer engineers $100,000 to tackle "commercially oriented" research topics from machine learning to data visualization and trading strategies, Goldman said Wednesday in a web post. While the company wants top students graduating in August to apply to the program, it's open to anyone.
I am traveling , Took a screen dump from some dismissal stats for a current project i am running since 2 days back on one of my 40 core machines. Thought it could be interesting to see. I am doing Fuzzy logic ones. They work pretty good . It is set to allow an entry if 70% of conditions is true for 5 conditions.
the issue i see with having a thread vs a subforum is that a single thread would not do well at keeping up with the vst amount of programs and topics, this would be covered by a subforum.
"Learning to Trade: The Cost Of Tuition"
- a roadmap of my lessons learned as taught by the market
What is most interesting about that paper is I am pretty sure it is utter nonsense to take 17 years worth of data as candlesticks, convert it to images and then use CNN.
Just like I have heard fast.ai Jeremy Howard say that support vector machines are basically useless vs ensemble methods. They gained in popularity because you could write a bunch of math about them and get a PhD. Just not reason to use one in practice though.
On the other hand for trading deep probabilistic programming has to have some use.
I am still trying to wrap my head around Pyro. Pyro