Deterministic Variable in Big Chaotic Financial Series

Today economists and especially market analysts need high-efficiency and robust tools for time series forecasting. Progress is rapidly stepping up, making traditional approaches less effective. The goal of this paper is to determine the origin of big financial series. In other words, do they contain deterministic variable component or fully chaotic? Is it possible to reconstruct certain process with the system of differential equations? How many equations do we need to fully explain the process behavior? A set of different time series was analyzed including currency rates, stocks, indexes and commodity prices. This paper is going to introduce several principles and approaches to help analysts extracting deterministic variables from highly noised time series. Especially this paper considers the time series, produced on financial markets. There were used two different algorithms to evaluating deterministic component of time series: functional method and Grassberg-Procaccia algorithm. The first approach is relatively easier to implement, than the second one, but result interpretation becomes a very difficult task, when data is noised. Thus, the major part of computations was performed using the Grassberg-Procaccia algorithm due to its high efficiency and interpretation friendly design.