Machine learning in trading: theory, models, practice and algo-trading - page 1625
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
Yu.I. Zhuravlev. Mathematical Methods of Forecasting
I have already posted it here... Very interesting lecture
I've posted it here before... Very interesting lecture
They say, if it were all data, let's do the math. I wonder if publicly available data would do, unemployment, oil stocks, and other things from the calendar.
They say, if they had all the data, we'll do the math. I wonder if publicly available data would work, unemployment, oil stocks and stuff from the calendar.
I don't think so.
Yu.I. Zhuravlev. Mathematical Methods of Forecasting
Thank you, this is an interesting report. Such approaches work well in the "man's game with nature" models, which is what the beginning of the report says. The market, for the most part, is a "people's game." The difference is in the nature of the uncertainty - in the first case it is "good" - probabilistic, and in the second it is "bad" - purely playful.
It is obvious that having reliable data it is possible to make sufficiently accurate predictions of certain phenomena. Other matter when it is a question of work at not full informativity. When a number of significant data is simply not available or not known at the moment. This is exactly what happens on the markets, when each participant works in his or her own field of informativeness about one or another quote. So... just thinking out loud...
This is true, but there is also a variant of banal criticism, for example, when a dataset of 50 samples)))
This is true, but there is also a variant of banal criticism, for example when dataset in 50 samples)))
Which describes 2 months of quotes excluding the noise you're all fighting with here. Yes, yes... we know... been there :-)
Or not, not that... That I'm Yura, right?
Or not. Let's feed the grid the minute data so it's on our watch.
Thank you, this is an interesting report. Similar approaches work well in "man's game with nature" models, which is what the beginning of the report talks about. The market, for the most part, is a "people's game." The difference is in the nature of uncertainty - in the first case it is "good" - probabilistic, and in the second it is "bad " - purely playful.
It seems to me that probability is probability in both cases...
The problem is the wrong experiment with the market when the experimental data is extracted, simply put, the statistics of deals or something like that...
The fractal structure of the market is not taken into account, nobody thinks about it, though it's obvious and explains a lot.
What everybody does is like standing by the sea and measuring water waves with a ruler, naively believing that the next wave will have the same size in centimeters)) nonsense.
Which describes 2 months of quotes excluding noise which you all are struggling with here. Yes, yes ... We know ... Been there :-)
Ah ..., well, if no noise, and then kick "vector machine Reshetov", then yes, it will be good, the main thing that was a lot of features, the more the ratio of features to the number of samples the cooler!
We will wait patiently for your stream!
Or not. Let's feed the grid with minute data so that it could trade on our watch. Ten thousand, and then we think, "Why don't we work?
Indeed, there are a lot of crazy people, ten thousand is just a blossom, some woodpeckers are trying to poke a million points! There are also tics and glasses...