Machine learning in trading: theory, models, practice and algo-trading - page 1000
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
I am not very familiar with the subject. I would like to understand - is it possible that detrending with ARFIMA will be useful for a sharp trend change (top or bottom)?
No.
We have to model all of the above. In particular, one may consider the problem of the subsequent model behavior after a spike.
The question of what they are doing here is more difficult to answer. Habit.
Another reason is probably because there is little choice, even on the fingers of one hand to count English forums with similar topics, where more than 1 post a day, one elitetrader and there too, liquidity has fallen, peak algotrader craze was somewhere in 2010, now people have realized that the ball is only cheese in a mousetrap, and turning analytical systems like those that recognize new particles in the LHC, it is for units, not for most. And here it's fun and diversified, like Babylon, like a primordial broth, but not degenerating into chaos thanks to moderation.
By the way, Mandelbrot's legacy is economophysics.
They have their own formulas and methods there, but I haven't studied them. It is postulated as a replacement for the obsolete theory of the efficient market.
The roots of economophysics lie in the works of the classics.Benoit Mandelbrot discovered in 1965 that the dynamics of financial series (price fluctuations in the stock exchange) are absolutely the same in small and large time scales: it's almost impossible to determine from the graph of such series, whether it represents price fluctuations during an hour, a day or a month. Mandelbrot called this propertyself-similarity, and the objects possessing it -fractals. Investigations of processes with such properties are carried out in physics with great energy, and analysis methods developed there often (unfortunately, not always) help to notice anomalies in behavior of financial series - the harbingers of sharp price collapses or rallies. The French mathematicianLouis Bachelier at the very beginning of the 20th century in his"Theory of Speculation" tried to describe the dynamics of financial series by analogy with Brownian motion - chaotic motion of molecules in a liquid or gas. Modern models which generalize such approach generate fractal processes which are statistically very similar to real financial series. Many of these models are based on the theory of chaotic dynamical systems - equations generating complexdynamics, sometimes almost indistinguishable from a random process - created in the 1970s-1990s. Modern econophysics also uses other powerful toolsof theoretical physics- for example, thecontinuum integral, the most important toolof quantum mechanics andquantum field theory. But perhaps the most fashionable trend today isevolutionary games, directly imitating the activities of countlessinvestors following one or another preference and principle.
Nowadays, an almost regular series of meetings on econophysics include: the NikkeiEconophysics Research seminar, and the APFA, ESHIA, Colloquium on Econophysics symposia.
The article in the English wiki seemed somewhat more meaningful to me. It seems that the main methods there are game theory and Monte Carlo simulation. My attitude to them is twofold: on the one hand, I partially agree with you and fxsaber's skepticism about Monte Carlo (in comments to my article), and on the other hand, I would be interested in simple game-based market models leading to non-stationary price series. Also interesting is that these methods can be a bridge between the technical and fundamental analysis. I can not insist that all this will necessarily help in trading, but it is possible to obtain some models, the parameters of which can be specified by the MO.
I read somewhere that game theory, until recently, had few applications in finance theory, but now there is progress. I would like to know more about it.
The article in the English wiki seemed to me somewhat more meaningful. It seems that the main methods there are game theory and Monte Carlo simulation. My attitude to them is twofold: on the one hand, I partially agree with your and fxsaber's skepticism about Monte Carlo (in comments to my article), and on the other hand, I would be interested to see simple game models of market leading to non-stationary price series. Also interesting is that these methods can be a bridge between the technical and fundamental analysis. I can not insist that all this will necessarily help in trading, but it is possible to obtain some models, the parameters of which can be specified by the MO.
I read somewhere that game theory, until recently, had few applications in finance theory, but now there is progress. I would like to know more about it.
For me, game theory for the market has developed in RL (the basics in my paper), where the payment matrix is replaced by a transition matrix or a parameterized stochastic agent policy. All this is relevant, of course, as long as the market strategy does not change. The basis there all the same is the fractal theory as applied to the market, in particular modeling through the Weierschrass-Mandelbrot function, as mentioned here just above and other analogues. I haven't tried to model these 2 together yet, but I have some thoughts on how to do interesting things. I haven't studied economophysics deeply and I don't know how it develops, judging by scarce information on the Internet - almost none :)
These are the graphs obtained after exponential thinning of the tick BP. As we can see, the variance is practically a constant both day and night.
For this purpose it is enough to take a window of dispersion calculation of the day size. No thinning has any effect here at all. If you knew how to do history tests, it would have been obvious a long time ago.
For me, game theory for the market is developed in RL (the basics in my article), where the payment matrix is replaced by a transition matrix or a parameterized stochastic agent policy. All this is relevant, of course, as long as the market strategy does not change. The basis there all the same is the fractal theory as applied to the market, in particular modeling through the Weierschrass-Mandelbrot function, as mentioned here just above and other analogues. I haven't tried to model these 2 together yet, but I have some thoughts on how to do interesting things. I haven't studied econophysics more deeply and I don't know how it develops now, judging by the scarce information on the Internet - almost none :)
RL is Reinforcement Learning?
It would be interesting to have game models directly related to the market. For example, you can try to simulate the process of hedging of traders' positions distortions by brokers. Maybe there are some stable patterns of price behavior (due to the inevitable time lag between the accumulation of skewness and its hedging). Although, everything must have been calculated a long time ago.
The English article doesn't have Mandelbrot at all for some reason. I may add it there.)