A0998
Title: Evaluating the statistical characteristics and analyzing the ML models of the limit order book
Authors: Dragana Radojicic - Faculty of Economics and Business, University of Belgrade, Belgrade, Serbia (Serbia) [presenting]
Abstract: The purpose is to describe the dynamics and statistical properties of the limit order book. The goal is to analyze the informativeness of features extracted from the limit order book. The basis is on real market data from the Nasdaq Stock Market. Firstly, relevant features are extracted from the database, and then their informativeness is analyzed. A customized data processing reconstructor is utilized to extract essential structures from the limit order book. Since there is potential in both analyzing historical stock data and developing algorithms to interpret that data. Different feature extraction methods and performances of the neural network models based on different topologies are examined.
