Machine learning in trading: theory, models, practice and algo-trading - page 1119
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
Trend = 100k lines. On the remaining 8k+(test) you apply the model. The data is shuffled.
Metric is logloss. Result posted. Trend =... test =...
I have not got much of a generalizing ability above 31 percent :-(.
Wizard, there is a variant to prepare 50 records actual that to trade models on OOS and put the result. I'll send you the model as I teach it???? Now straight!!!!
I think there are fish in this hundred lines, because the algorithm does not fall out, but stubbornly uses a large number of inputs.....
So there is an option to run my models on your data????
Sensei, thanks for the vote of confidence, but I do not need anything)))
So it's just for fun....
Amazing... I reduced the sample to 40 records in the end the training results remained at the same level, where 30% on your data compared to 70% on mine, which shows that my data is much better than yours and it surprised me if you earn them how?????
Trend = 100k lines. On the remaining 8k+(test) you apply the model. The data is shuffled.
Logloss metric. Result posted. Trend =... test =...
DateTime why did you cut it out... mixed it up... It's a time series, not some bank scoring...
)))))))))))))))) Another episode of the Marlesonian ballet, by Mikhail))) It's nice that in the world of variability and chaos there are islands of such unshakable stability))))
And a year ago, Mikhail even promised to master Java, but...
So I'm doing it for a little bit.... Java is a thing :-)
I didn't cut or mix anything. You're the one who's been complaining. Do it.
In the end your data sucks compared to mine and if you trade with it I pity you :-(
I'm tired of laughing today)))
Well, you betcha. Throw the guys a sentimental set of "what the hell" and sit there laughing as people are trying to turn shit into candy, not good with comrades in arms, oh how not good....
Not the kid, the kid. I don't know what the hell you're doing.)
Aah, you're giving him a place in our community. Nice plan. Let's see what he and his AI can do :-)
This case, and with your dataset... sorry, but a lot of people told you many times that you need at least thousands of samples, and given the noisiness of market data preferably hundreds of thousands of points, but when you learn Java and use XGB for example, you'll laugh at your persistence in the past))
Dude well you yourself understand for my TS 1000 points is 50 weeks of time. I'm not sure that on such a stretch you can get a good quality model without overtraining, if it's not some super giant AI from google, as an example ....
I didn't cut or mix anything. I resented it. Do it.
What are your results? So that there is some interest.
So I came up with an improved metric for the Matthews coefficient, but what can I say if you blow here and blow there? :-(
I'm stuck with passing an array from one class to another, and I'm screwed... no one to ask :-(