Development of a cyber-physical system in Python and QLua for trading on the QUIK platform on MoEx in line with the digitalization of the economy
- 18 March 2024
- journal article
- research article
- Published by PANORAMA Publishing House
- No. 3,p. 214-231
Abstract
The article discusses theoretical approaches to modeling stock trading robots — cyber-physical systems in the conditions of innovative transformations and large-scale digitalization of the economy and financial sector. The relevance of the study is that currently there is a tendency to increase the share of algorithmic trading on the stock exchange, and the share of trading robots based on artificial intelligence is increasing. A fresh technological trend is the use of stock trading robots — cyber-physical systems based on cognitive modeling, artificial intelligence algorithms operating in digital ecosystems. Robotic algorithms compete with each other, tracking the dynamics of submitted orders, looking for "densities" in the order book, changing the frequency of purchase/sale transactions, monitoring the entire market of financial instruments, tracking spikes in volatility, catching transactions of large players in the table of impersonal transactions, adjusting the parameters in your scripts online. The scientific novelty lies in the fact that in the presented study a deep learning model DL model "Random Forest" was formed, which calculates the forecast of the closing price of the SiZ3 futures contract on the required time frame. The practical signifi cance of the study is that the results obtained have been implemented and are actively used in stock trading. The criterion for the success of the predictive properties of the DL model was the value of the average forecast error (MAE). The proposed DL model uses the best decision tree, which has optimal hyperparameter settings, for example, the depth of the tree is six layers, the number of estimators (trees) in the ensemble is ten. In the experiment, the hyperparameters of the neural network did not change; the input parameters to various trees were selected randomly by the algorithm. The DL model showed high forecast accuracy.Keywords
This publication has 13 references indexed in Scilit:
- Improving Accuracy and Reducing Financial Risk When Forecasting Time Series of SIU0 Future Contracts Employing Neural Network with Word2vec Vector News2022
- Algorithmic Trading Efficiency and its Impact on Market-QualityAsia-Pacific Financial Markets, 2021
- Complex Monitoring Systems for Landfills2021
- Development of Financial Service Methods for People with Dementia during Digitalization: A Partnership between Citizens and the Russian StateInternational Journal of Technology, 2020
- Features of segmentation and positioning processes when creating an educational brand in neural network economy2020
- Analysis of Factors, Defining Software Development Approach2017
- Small innovative company’s valuation within venture capital financing of projects in the construction industryMATEC Web of Conferences, 2017
- The role of algorithmic trading systems on stock market efficiencyInformation Systems Frontiers, 2013
- The New Secondary Market Structure: Competition, Dark Pools, Algorithmic and High-Frequency Trading2012
- Risk, Return, and Equilibrium: Empirical TestsJournal of Political Economy, 1973