[lyft]

Let me preface this by saying I'm not a crypto-currency expert. In fact, I don't even trade. With that being said, I'm a pretty strong developer/ data scientist and can build scale-able machine learning models. I'm not sure what the use would be, but I ran tests with stocks before and some stock graphs tested with 96%+ confidence with my models before. Obviously these are stocks that have had consistent growth, but they can be moderately accurate.

I could build an API with node and give you guys endpoints (selling cheap ass access keys), or I could just throw this up on a website, or a discord bot. It doesnt matter to me.

Obviously it would be extremely cheap, as its not going to make you the next Warren Buffet; however, when it comes to AI/ machine learning, sometimes there are patterns and trends that the human eye cannot catch. All forecasts would show a certain number of days out with prices and also the confidence.

So say you wanted to run a prediction on Bitcoin 15 days out, it might say '6120$ @ 87% confidence' where the confidence is how well this graph fits my machine learning models.

I would also allow users to choose the machine learning algorithms. If anyone is interested in this, or even the math/code behind it, let me know. Im just a dude excited about AI in the finance game.

---

edit: Honestly with something like this it's more novelty than anything, but I wouldnt doubt some people can make some money with this

[Golang]

Damn, that sounds fucking awesome!

I wish I knew how to make something like this work, I feel like with enough training data this could make something kinda profitable.

[WillyT1998]

I am new to the bitcoin market and I would love to know when to sell out. With that said, this will be a huge benifit for someone like me who is just starting out. I would love to buy an access key if you get it going, please pm me if you are interested in my offer.

[kanibaL]

Why wouldnt you get it testing for some longer time, let it learn a lot, and when it does, you could prove that it can really forecast and then create some kind of platform that you could sell multiple times for really serious money.

[lyft]

Why wouldnt you get it testing for some longer time, let it learn a lot, and when it does, you could prove that it can really forecast and then create some kind of platform that you could sell multiple times for really serious money.

That's what I'm trying to do! I think it would work well with some neural evolution algorithms that are floating around out there. It could learn from mistakes and reduce error as it goes along. I know for a fact that some models I've worked with before would perform well on bitcoin.

[Ninetales.]

Check out Gunbot you may have a bit to copy from over there.

[kanibaL]

That's what I'm trying to do! I think it would work well with some neural evolution algorithms that are floating around out there. It could learn from mistakes and reduce error as it goes along. I know for a fact that some models I've worked with before would perform well on bitcoin.

And the Bitcoin trading is just a fragment as there is a loooot of popular cryptocurrencies other than Bitcoin, not even to metion the other kinds of trading/forex/stocks and such things.

[lyft]

Check out Gunbot you may have a bit to copy from over there.

Thank you sir! Wow, they are pricey. I wonder if they are using machine learning or they just hard code some signals that the program looks for.

[th3c0rrupt3d0n3]

Thank you sir! Wow, they are pricey. I wonder if they are using machine learning or they just hard code some signals that the program looks for.

Gunbot is nothing close to machine learning. Is actually a nice tool to make money on bullish markets, but wow, who cant make money on bullish markets.

Good luck with this bro, sounds like a nice project id love to see.

[BitcoinLuxury]

If you make this please let me know I would be interested, i am a big fan of these kind of data tools.