Machine learning in trading: theory, models, practice and algo-trading - page 964
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
What else would you say is interesting?
Then you shouldn't have shed a tear, becausethe algleeb itself doesn't let you see anything else . It is known for a long time, except for you.
I've got it all working, no one has ever shown me better charts
I wrote my own framework and work with it
If I can think of a way to use p or python, I will do it.
Oh watch out, Maximka, taking investors' money. You need a picture like this for this kind of money and I wouldn't risk taking it. Just some friendly advice... don't risk it. Your stats are not good. It's really not clear when TC will start to drain, and this moment is one of the IMPORTANT. You think you have it in a drawdown, but it's been draining for a long time..... That's a lot of money. Then you'll have a long time to pay it back.....
With my stats, I don't dare take it now. It's better to pile up in the volume of trades and earn good money on your own. When the lot is dynamic the deposit may grow exponentially fast. The main thing is not to make mistakes.
With your stats I wouldn't dare either ))
I'll have to python and get a shorter model, most likely, but otherwise i'll be fine.
let's use xgboost instead of scaffolding
You can't be in the market without insiders and connections.
You finally got it! I taught you this a year ago, and you're all neural networks, matlab, python... Who was right in the end? Listen to what I'm saying, first learn how to do it right, and only then experiment.
You finally got it! I taught you this a year ago, and you're all neural networks, matlab, python... Well, who was right in the end? Listen to what I'm saying, first learn how to do it right, and only then experiment.
Ohhhhhh the stableman hatched ))))
You finally got it! I taught you this a year ago, and you're all neural networks, matlab, python... Well, who was right in the end? Listen to what I'm saying, first learn how to do it right, and only then experiment.
So show me how to do it right??? Where are your statistics??? And then we'll talk....
With your stats I wouldn't dare either ))
I'll have to python and get a shorter model, most likely, but that's ok
let's go with xgboost instead of scaffolding.
What's wrong with it? Especially with the last screen. There are few deals, but the quality of them is amazing.... No slippage at all. Not like you... Isn't it?
What didn't you like about her? Especially from the last screen. There are not many deals, but the quality of the deals is amazing.... No slippage at all. Not like you... Isn't there?
Not many, for the quadrillionth time... I can do that too.
and my screenshots show 4 months of training and almost a year of OOS and some show more
and, by the way, my deck didn't leak - these are model tests. I'll lock it later again later on. Need to deal with the models, can't figure out if they can be improved or not yet. I will transfer them to python and work on them some more... I wanted to use P at first but then I remembered it pisses me off
little, quadrillionth time I say... I can do these too
and my screenshots show 4 months of training and almost a year of OOS, and some show more
And, by the way, my deck did not merge - it's model tests. I'll lock it later again later. Need to figure out the models, can't figure out if they can be improved or not yet
Anyway, you do not take other people's money. They'll put you on the hook later. You'll know.... :-)))
Then why try to filter out the bad and noisy examples; or isolate them, repartition them to "don't know" and train the network again?
The question is correct, the answer is trivial - fitting a matrix model that does not match the input data.
I've been reading such threads for years and constantly see the phrase about noise - why is there noise in the price series? - When it comes to Open and Close - well, the time is over on the timeframe, the price is fixed at the moment. When it comes to High and Low - well, someone was lucky to place an order on the very "pip" (or maybe not) - what noise? Every price movement is a real market participant action, even OC with its smoothing of price movements (tick filtering) - also a market participant.
ZZZY: it is possible to filter the input data, use large TFs - the filter is justified and logical, filter the sections with high volatility and low - the filter is justified and logical, filter the trading time - the filter is logical and justified.... And discard the input data because it's noise... well that's probably an option too, how reasonable is it? - ah yes the math model requires it so the chart in the tester is nice ))))