Forecasting high‐frequency excess stock returns via data analytics and machine learning

Abstract

Borsa Istanbul introduced data analytics to present additional information about its market conditions. We examine whether this product can be utilized via various machine learning methods to predict intraday excess returns. Accordingly, these analytics provide significant prediction ratios above 50% with ideal profit ratios that can reach up to 33%. Among all the methods considered, XGBoost (logistic regression) performs better in predicting excess returns in the long‐term analysis (short‐term analysis). Results provide evidence for the benefits of both the analytics and the machine learning methods and raise further discussion on the semistrong market efficiency.

Machine learning methods in finance: Recent applications and prospects

“…Chinco et al (2019) apply LASSO to predict ultra-short-term future stock returns based on the cross-section of ultrashort-term historical returns. Akyildirim et al (2021) Most studies only focus on the predictions themselves. However, there are also some studies that try to analyse how the predictor variables affect the predictions.…”

“…Chinco et al (2019) apply LASSO to predict ultra‐short‐term future stock returns based on the cross‐section of ultrashort‐term historical returns. Akyildirim et al (2021) use various ML methods to predict intraday excess returns based on high‐frequency order and trade information. Amel‐Zadeh et al (2020) predict abnormal stock returns around earnings announcements based on financial statement variables.…”

“…construct a hybrid approach that combines traditional methods based on financial theory with ML to predict future excess stock returns Chinco, Clark-Joseph, and Ye (2019). apply LASSO to predict ultra-short-term future stock returns based on the crosssection of ultrashort-term historical returns Akyildirim et al (2021). use various ML methods to predict intraday excess returns based on high-frequency order and trade information.…”

Modelling failure rates with machine‐learning models: Evidence from a panel of UK firms

“…Our study makes three important contributions to the literature. First, we rely on machine learning (ML) tools, which are gaining ground in economic and finance applications (see, e.g., Akyildirim et al, 2021; Athey et al, 2019; Aziz et al, 2021; Barboza et al, 2017; Knaus et al, 2021; de Moor et al, 2018). We compare the performance of these models with the golden standard of default prediction studies—the discrete hazard (DH) model.…”