|
|
Výběr délky regresoru
Křivánek, O. ; Zeman, Jan
This research report is closely connected to the long time running research of the usage of the theory of Bayesian learning, stochastic dynamic programming and its approximations in futures dealing problem. This report describes tuning of one selectable parameter, which occurs in the new-designed algortihm called iterations-spread-in-time strategy. Experiment is done on real economic data on 35 selected futures markets. The main criterion of succes is the so-called net profit and also comparison with the previous experiments.
|
|
|
Testování zapomínacího faktoru
Votava, A. ; Zeman, Jan
Presented work deals with forgetting estimation in the frame of dynamic decision making. The main goal is to find the optimal forgetting system for the algorithm for estimation of forgetting factor in time in the optimal way. Further goal is to compare the algorithm with the constant forgetting for various settings.
|
|
|
Nový přístup k odhadování Bellmanových funkcí
Zeman, Jan
The paper concerns an approximate dynamic programming. It deals with a class of tasks, where the optimal strategy on a shorter horizon is close to the global optimal strategy. This property leads to a new, specific, design of the Bellman function estimation. The paper introduces the proposed approach and provides an illustrative example performed on the futures trading data.
|
|
|
Konstrukce vícekrokových predikcí v normálním BVAR(p) modelu za použití Monte Carlo vzorkování
Šindelář, Jan
In Bayesian normal vector AR model (BVAR) of data evolution in a discrete time we are trying to predict the distribution of data up to horizon h. Since the analytical solution of such a prediction is difficult due to the high dimensionality of the problem, we are forced to search for approximative solutions. We propose a solution using Monte Carlo sampling from parameter distribution and later reconstruction of the predictive distribution of data.
|
|
|
Parciální zapomínání. Nová metoda sledování časově proměnných parametrů
Dedecius, Kamil ; Nagy, Ivan ; Kárný, Miroslav ; Pavelková, Lenka
Tracking of slowly varying parameters is an important task in the theory of adaptive systems. Majority of prediction and control algorithms, employing regression models like autoregression model (AR), autoregression model with exogenous inputs (ARX), autoregression model with moving average (ARMA) etc., assume a carefully defined model structure and correctly estimated parameters. Problems arise, when the model parameters vary in time. The problems of slowly time-varying model parameters were given a thorough attention. The proposed partial forgetting method tries to solve this issue by a new approach.
|
|
|
Vylešpení modelu dynamického rozhodování pomocí metody "Iteration spread in time"
Divišová, L. ; Zeman, Jan
In the present work we study the problem of ¯nding the best de- cision based on our previous experience with the system. To solve this task, we use the dynamic programming and its approximations. In the work we summarize the theory needed for usage of the dynamic programming and we deal with its application on futures dealing trying to ¯nd best strategy, id est a sequence of decisions, maximizing our gain or minimizing the loss function. We introduce notion "Bellman function", explain why the approximation of this function is needed, demonstrate one of already tested approximation methods together with its results and we try to propose a method that would lead to the best approximation in suitable time and with available computation aids.
|
|
|
Zpřesnění modelu odhadování vývoje cen na komoditních trzích za pomocipřidání nových kanálů
Kozmík, V. ; Šindelář, Jan
Present work deals with a problem of price prediction on futures markets. Main goal of this work is to find out on which input channels (price,volume of contracts, etc.) depends the sought future price. We make further simplications to be able to present the problem as a mathematical one, which can be solved using Bayesian estimation methods. In this work we present the process of price prediction and then concentrate on the main aim, which is structure determination. Applied algorithm is described mathematically and also programmed in Matlab environment. The results of the algorithm on specific data and their economic interpretation are the last part of this work.
|
|
|
Implicitní aproximace Bellmanovy rovnice
Pištěk, Miroslav
In this article, an efficient algorithm for an optimal decision strategy approximation is introduced. It approximate the Bellman equation without omitting the principial uncertainty stemming from an uncomplete knowledge. An integral part of the proposed solution is a reduction of memory demands using HDMR approximation. The result of this method is a linear algebraic system for an approximated upper bound on the Bellman function. One illustrative example has been completely resolved.
|
|
|
Návrh strategie pro obchodování s futures kontrakty
Zeman, Jan
The paper considers by design of trading strategy for futures markets. The problem is defined as dynamic decision making task and it is solved using iteration spread in time and Monte Carlo method. The paper includes results on experiments with real data.
|
|
|
Dynamic decision making via approximate dynamic programming
Slimáček, V. ; Zeman, J. ; Kárný, Miroslav
This work deals with dynamic decision making via approximate dynamic programming in application to the futures trading. This work describes the theoretical description of dynamic decision making and approximate dynamic programming, you can also nd here the principles of Bayesian estimation, which is necessary for solving our task.We designed and described one of possible trading strategy { receding horizont strategy combined with anticipative strategy, the predictions of prices, which are necessary for using of this strategy, are made by certainty equivalence strategy and by Monte Carlo method. The designed strategy was tested on real data and unfortunately this strategy doesn t provide a prot.
|
guest ::
