Intraday Seasonality in Analysis of UHF Financial Data: Models and Their Empirical Verification

Abstract

The aim of this paper is to outline the typical characteristics of the ultra-high-frequency financial data and to present estimation methods of intraday seasonality of trading activity. Ultra-high-frequency financial data (transactions data or tick-by-tick data) is defined to be a full record of transactions and their associated characteristics. We consider two nonparametric estimation methods: cubic splines and a Nadaraya-Watson kernel estimator of regression. Both approaches are compared empirically and applied to financial data of stocks traded at the Warsaw Stock Exchange.