SEARCH WITHIN CONTENT
JEL classification : C53, G17
Citation Information : Statistics in Transition New Series. Volume 18, Issue 1, Pages 91-114, DOI: https://doi.org/10.21307/stattrans-2016-059
License : (CC BY 4.0)
Published Online: 07-July-2017
The main goal of this paper is to present the method for describing and predicting trade intensity on the Warsaw Stock Exchange. The approach is based on generalized linear models, the variable selection is performed using shrinkage methods such as the Lasso or Ridge regression. The variable under investigation is the number of trades of a particular stock 5-minute interval.The main conclusion is that the number of trades during short intervals is predictable in the sense that the prediction, even based on relatively simple models, is with respect to statistical properties better than the prediction based on the random walk, which is used as a benchmark model.