Pages:
Author

Topic: Beta version of new Bitcoin price prediction model! - page 2. (Read 3877 times)

sr. member
Activity: 322
Merit: 250
Got a neural network to tell folks which neural network they should follow at any particular moment?  Looks like they give conflicting calls.


The 24 hour one - it has the lowest average error. The 20 day one is unreliable, as it says above the chart. The 5 day one does OKAY but it's not as good (as is reflected in the higher average error)
sr. member
Activity: 364
Merit: 250
Got a neural network to tell folks which neural network they should follow at any particular moment?  Looks like they give conflicting calls.
sr. member
Activity: 322
Merit: 250
Yeah the 24 hour predictor on the home page has been doing really well recently... I'm almost tempted to just scratch the new predictor. The cool thing about the way this works is that the longer I keep it running, the more accurate it will become. As more data becomes available to train the neural network, it can learn more and the average error becomes lower.
hero member
Activity: 552
Merit: 501
I don't much believe in TA, whether it's done by a human or a computer. But I must admit I'm impressed by the prediction made eight hours ago of a fall from 670 to 635 which was almost spot on both in terms of timing and degree.

Were you looking at the chart here:
http://www.btcpredictions.com/Historic.php

or here?
http://www.btcpredictions.com/NewPredictor.php

For a few hours after i posted last, I checked in with the NewPredictor every once in a while, which said we'd be going through a rising swell today. The whales apparently weren't going to have that happen today. The insane thing is that the supposedly simpler predictor at Historic.php forsaw the crash 6 hours before it happened.

Got me sitting up straight and watching intently now.....

The historic (3 layer once an hour). Uncanny.
sr. member
Activity: 303
Merit: 250
I don't much believe in TA, whether it's done by a human or a computer. But I must admit I'm impressed by the prediction made eight hours ago of a fall from 670 to 635 which was almost spot on both in terms of timing and degree.

Were you looking at the chart here:
http://www.btcpredictions.com/Historic.php

or here?
http://www.btcpredictions.com/NewPredictor.php

For a few hours after i posted last, I checked in with the NewPredictor every once in a while, which said we'd be going through a rising swell today. The whales apparently weren't going to have that happen today. The insane thing is that the supposedly simpler predictor at Historic.php forsaw the crash 6 hours before it happened.

Got me sitting up straight and watching intently now.....
hero member
Activity: 552
Merit: 501
I don't much believe in TA, whether it's done by a human or a computer. But I must admit I'm impressed by the prediction made eight hours ago of a fall from 670 to 635 which was almost spot on both in terms of timing and degree.
sr. member
Activity: 303
Merit: 250
Excellent work. I've been reading The Singularity Is Near (Kurzweil) so anything neural network piques my interest  Smiley
full member
Activity: 238
Merit: 100
Also in the little test you ran, how many Bitcoins did you have on Nov 30 2013 and April 10 2014

On Nov 30 2013 it had 86.290957 BTC. It's really hard for me to tell how much it had on April 10 2014 but it looks like probably around 127 ish

Awesome stuff. Let's see how your new beta predictor (on the site) plays out today.
sr. member
Activity: 322
Merit: 250
Also in the little test you ran, how many Bitcoins did you have on Nov 30 2013 and April 10 2014

On Nov 30 2013 it had 86.290957 BTC. It's really hard for me to tell how much it had on April 10 2014 but it looks like probably around 127 ish
full member
Activity: 238
Merit: 100
This looks great...an amazing creation you have done here.
How long did it take to create?

Glad you like it Smiley

I've been working on it since around October or November 2013, but a lot of that was also just learning about Neural Networks. When I started, I really didn't have a thorough understanding of them, just a pretty basic understanding. Took a while to learn. Also spent some time learning php/javascript/html so that I could make the website.

Also in the little test you ran, how many Bitcoins did you have on Nov 30 2013 and April 10 2014
full member
Activity: 238
Merit: 100
Did you do the recent test with the new beta?

Btw, sent you a donation

Thanks a lot for the donation, I really appreciate it!

Is the test you are asking about the one that I said earned $60k? That was done using a different type of neural network, not the beta of the 24 hour one that is up now. It was done with a neural network I've been working on that predicts the highest and lowest price during the next 24 hour period rather than every price.

Keep working on that, I see real applications for it and I would be interested in supporting a more developed version.
sr. member
Activity: 322
Merit: 250
This looks great...an amazing creation you have done here.
How long did it take to create?

Glad you like it Smiley

I've been working on it since around October or November 2013, but a lot of that was also just learning about Neural Networks. When I started, I really didn't have a thorough understanding of them, just a pretty basic understanding. Took a while to learn. Also spent some time learning php/javascript/html so that I could make the website.
member
Activity: 70
Merit: 10
This looks great...an amazing creation you have done here.
How long did it take to create?
sr. member
Activity: 322
Merit: 250
Did you do the recent test with the new beta?

Btw, sent you a donation

Thanks a lot for the donation, I really appreciate it!

Is the test you are asking about the one that I said earned $60k? That was done using a different type of neural network, not the beta of the 24 hour one that is up now. It was done with a neural network I've been working on that predicts the highest and lowest price during the next 24 hour period rather than every price.
full member
Activity: 238
Merit: 100
uh, it's very bearish  Cry

Yeah to be honest, the 24 hour prediction on the main page is the only one that I would say is actually pretty good. The 5-day one is good sometimes but iffy overall, and the 20 day one is really just experimental.

Will you add longer predictions soon?

I am working on trying to get the new predictor to predict prices over a longer period of time. I am also working on the buy/sell price predictor though. And I'm also still working on the 24 hour new prediction, because it doesn't seem to be changing shape very much and I'm not really sure why yet. Ahhhh so many things to work on!

Did you do the recent test with the new beta?

Btw, sent you a donation
sr. member
Activity: 322
Merit: 250
uh, it's very bearish  Cry

Yeah to be honest, the 24 hour prediction on the main page is the only one that I would say is actually pretty good. The 5-day one is good sometimes but iffy overall, and the 20 day one is really just experimental.

Will you add longer predictions soon?

I am working on trying to get the new predictor to predict prices over a longer period of time. I am also working on the buy/sell price predictor though. And I'm also still working on the 24 hour new prediction, because it doesn't seem to be changing shape very much and I'm not really sure why yet. Ahhhh so many things to work on!
legendary
Activity: 2772
Merit: 1028
Duelbits.com
uh, it's very bearish  Cry

You're looking at the old version.

New version is short term, but very bullish

http://www.btcpredictions.com/NewPredictor.php

yeah, just noticed it  Grin

Will you add longer predictions soon?
sr. member
Activity: 322
Merit: 250
Would be really sweet if you could do some simulations during the 6 month 'depression' we were in. I think that would get many people interested in your little cool project, including me.

I'll work on that, that shouldn't be too hard to do.

Quote from: Raystonn
Did you train this on all available data, or did you leave some for forward testing?  You need to watch for curve fitting.

It trains on all of the data. I don't think it's overfitting because it actually has a relatively high average error by the time it's one training (something around 8%) but it's still able to profit. I'll definitely be testing it out so we'll see how it performs. Maybe I'll put it online so people can track its progress.
full member
Activity: 238
Merit: 100
uh, it's very bearish  Cry

You're looking at the old version.

New version is short term, but very bullish

http://www.btcpredictions.com/NewPredictor.php
legendary
Activity: 2772
Merit: 1028
Duelbits.com
uh, it's very bearish  Cry
Pages:
Jump to: