A novel feature engineering approach for high-frequency financial data
Novel feature engineering for high-frequency financial data using time series segmentation to forecast trends.
What it examines
This paper proposes a novel feature engineering approach for high-frequency financial data using time series segmentation. The methodology extracts and analyzes variables by intraday trends to forecast future trends, applied to estimate high-frequency volatility, duration, and direction using XGBoost on data from the Brazil Stock Exchange.
What it concludes
The research demonstrates the potential of the proposed feature engineering methodology for high-frequency financial data analysis and forecasting. Potential applications include improved financial market predictions and risk management. Future research could explore broader applications and refine the methodology for enhanced accuracy.
Evidence objects
The research demonstrates the potential of the proposed feature engineering methodology for high-frequency financial data analysis and forecasting. Potential applications include improved financial market predictions and risk management. Future research could explore broader applications and refine the methodology for enhanced accuracy.
key_findings bullet 1 · key_findings · validation V0
Raw abstract and provenance
Highlights A novel feature engineering approach for high frequency financial data Pablo Mantilla,Sebasti?n Dormido-Canto ?It is proposed a feature engineering for high frequency financial data based on time series segmentation. ?This methodology allows to extract and analyze variables by intraday trends, as well as feeding artificial intelligence models in order to forecast response variables in future trends. ?Thisfeatureengineeringisappliedtoestimatehighfrequencyvolatility,durationanddirectionlinkedtofutureintraday trends. ?Experimentation was conducted using high frequency financial data from the Brazil Stock Exchange. A novel feature engineering approach for high frequency financial data Pablo Mantilla, Sebasti?n Dormido-Canto Universidad Nacional de Educaci?n a Distancia (UNED), Madrid, Spain ARTICLE INFO Keywords : Artificial intelligence in finance High frequency financial data Time series segmentation Intraday volatility Directional forecastingABSTRACT It is proposed a feature
Source row: 68 · abstract type: unknown