Realistically speaking, is it possible to predict stock market prices using ML? Isnt the data to random and broad to make a model out of it? The past data wouldnt relate to the future data (because of said randomness) rendering the whole model useless. Example: You feed all the past data to a model to predict future cadles, if the predition is wrong, the model would have to be refed all past data for next prediction and the process continues to inifinity, making the model no more accurate than flipping a coin and less efficient than looking at a weekly/daily chart making and making conclusion yourself based on human logic.
Idk if what i said makes sense and if my understanding of ML is right. >inb4 ESL
after the Citibank drama last week I doubt it's a good idea
Jayden Reyes
yes, it's called being a quant if your only input is previous price data then you're just doing TA, might as well invest in a deck of tarot cards instead
Juan Peterson
Imagine that thing slithering in top of you and grunting while it ruts with your body
Xavier Smith
The problem with machines are, they lack common sense. It would make sense if you combine all the datapoint like orderbook, liquidity, news analysis, volume, indicators, history of success on indicators, all time frames etc) which would consume so much computing power and resources you would need at least 1 supercomputer.
>You feed all the past data to a model to predict future cadles Sounds like a gambler's fallacy machine. You'll probably lose big. People who make money on stocks do so because they have access to information that others don't, or gamble blind and hit
David King
>People who make money on stocks do so because they have access to information that others don't like a Bloomberg terminal
Jose Smith
no it isn't. yes it's random.
Jack Rivera
Link? Isnt quant just automated TA, and hedge funds trusting the math backing it up? Quats are also the guys making high frequency trading software (most of the time), which has no risks involved. >TA is a meme Controversial topic. On small time frames (less than 4h, yes)
Aaron Cook
>aka your iq is sub 80 Did you even read a single word I wrote? t.125 IQ
Joseph Price
>Realistically speaking, is it possible to predict stock market prices using ML? if your neural network has access to all of the insider trading info, then yes. though, you probably wouldnt want to bother with ML at that point
Matthew Gray
>humans can learn it
they literally can't. stock fluctuations are random.
Gabriel Clark
I tried to identify patterns, but then read online that PowerBall and Mega Millions both have multiple sets of balls that are picked at random, so my hope that some imperfection in a ball caused it to hit more or less frequently went out the window quick.
Eli Davis
Wow, you're probably the first person that has ever thought of doing this.
Wyatt Hill
I’ve made a living for the past ~2 years doing this, but if you have to ask then you individually couldn’t, no, no offense.