Grid Trading System Robot (GTSbot): A Novel Mathematical Algorithm for Trading FX Market

Abstract

Grid algorithmic trading has become quite popular among traders because it shows several advantages with respect to similar approaches. Basically, a grid trading strategy is a method that seeks to make profit on the market movements of the underlying financial instrument by positioning buy and sell orders properly time-spaced (grid distance). The main advantage of the grid trading strategy is the financial sustainability of the algorithm because it provides a robust way to mediate losses in financial transactions even though this also means very complicated trades management algorithm. For these reasons, grid trading is certainly one of the best approaches to be used in high frequency trading (HFT) strategies. Due to the high level of unpredictability of the financial markets, many investment funds and institutional traders are opting for the HFT (high frequency trading) systems, which allow them to obtain high performance due to the large number of financial transactions executed in the short-term timeframe. The combination of HFT strategies with the use of machine learning methods for the financial time series forecast, has significantly improved the capability and overall performance of the modern automated trading systems. Taking this into account, the authors propose an automatic HFT grid trading system that operates in the FOREX (foreign exchange) market. The performance of the proposed algorithm together with the reduced drawdown confirmed the effectiveness and robustness of the proposed approach.

Deep LSTM with Dynamic Time Warping Processing Framework: A Novel Advanced Algorithm with Biosensor System for an Efficient Car-Driver Recognition

“…At this point, the PPG signals coming from the various sensing devices placed on the steering were suitably filtered and stabilized, and ready to be used to characterize the driver's cardiovascular imprint. Next, a further preprocessing of the collected PPG time series was proposed using dynamic time warping (DTW) techniques [19,20] with respect to a PPG reference signal generated by ad hoc reaction-diffusion model (RD-M) [11] which was able to generate fully compliant PPG patterns. It is The collected raw PPG time series coming from the PPG sensing devices, were further post-processed by means of ad hoc dynamic time warping (DTW) processing (made by the DTW block) followed by deep learning blocks, respectively.…”

“…It is well known that two time-varying signals can have similar features but appear very different because the temporal distribution of the features is different in each signal. The DTW technique made it possibe to create distortions in the temporal distribution of the features of a signal with respect to one considered reference until a distribution that maximized the temporal similarity (defined by specific metric) of the two signals, was found [20]. Formally, given two K-dimensional signals (ζ, ϑ):…”

“…Among the possible adoptable, the symmetric Kullback-Leibler (SKL) metric (see Equation (4)) was proposed because, in our tests, it showed better performance than the classic Euclidean or absolute. More details about DTW and SKL metric are available in [19,20].…”

“…..n; x = (HP, LP); y = (HP, LP) ∀ n, m ∈ N + (14) Figure 8 shows an example of the DTW processing through the SKL metric between a signal patterns of the car-driver PPG signal (coming from biosensing device on the steering) and the corresponding signal pattern generated from the reference RD-M: Figure 8 shows an example of the DTW processing through the SKL metric between a signal patterns of the car-driver PPG signal (coming from biosensing device on the steering) and the corresponding signal pattern generated from the reference RD-M: The collected setup of similarity values ( ) are used as input for the deep long short-term memory (LSTM) model [20][21][22][23][24] described in the next subsection.…”

“…The proposed deep learning pipeline is depicted in Figure 9. The collected setup of similarity values δ k xy (k) are used as input for the deep long short-term memory (LSTM) model [20][21][22][23][24] described in the next subsection.…”