Showing 1–4 of 4 results for author: Dupuis, A

Improving predictability of time series using maximum entropy methods

Authors: Gregor Chliamovitch, Alexandre Dupuis, Bastien Chopard, Anton Golub

Abstract: We discuss how maximum entropy methods may be applied to the reconstruction of Markov processes underlying empirical time series and compare this approach to usual frequency sampling. It is shown that, at least in low dimension, there exists a subset of the space of stochastic matrices for which the MaxEnt method is more efficient than sampling, in the sense that shorter historical samples have to… ▽ More We discuss how maximum entropy methods may be applied to the reconstruction of Markov processes underlying empirical time series and compare this approach to usual frequency sampling. It is shown that, at least in low dimension, there exists a subset of the space of stochastic matrices for which the MaxEnt method is more efficient than sampling, in the sense that shorter historical samples have to be considered to reach the same accuracy. Considering short samples is of particular interest when modelling smoothly non-stationary processes, for then it provides, under some conditions, a powerful forecasting tool. The method is illustrated for a discretized empirical series of exchange rates. △ Less

Submitted 28 November, 2014; originally announced November 2014.

Comments: 4 pages, 4 figures

Multi-scale Representation of High Frequency Market Liquidity

Authors: Anton Golub, Gregor Chliamovitch, Alexandre Dupuis, Bastien Chopard

Abstract: We introduce an event based framework of directional changes and overshoots to map continuous financial data into the so-called Intrinsic Network - a state based discretisation of intrinsically dissected time series. Defining a method for state contraction of Intrinsic Network, we show that it has a consistent hierarchical structure that allows for multi-scale analysis of financial data. We define… ▽ More We introduce an event based framework of directional changes and overshoots to map continuous financial data into the so-called Intrinsic Network - a state based discretisation of intrinsically dissected time series. Defining a method for state contraction of Intrinsic Network, we show that it has a consistent hierarchical structure that allows for multi-scale analysis of financial data. We define an information theoretic measurement termed Liquidity that characterises the unlikeliness of price trajectories and argue that the new metric has the ability to detect and predict stress in financial markets. We show empirical examples within the Foreign Exchange market where the new measure not only quantifies liquidity but also acts as an early warning signal. △ Less

Submitted 10 February, 2014; originally announced February 2014.

The scale of market quakes

Authors: T. Bisig, A. Dupuis, V. Impagliazzo, R. B. Olsen

Abstract: We define a methodology to quantify market activity on a 24 hour basis by defining a scale, the so-called scale of market quakes (SMQ). The SMQ is designed within a framework where we analyse the dynamics of excess price moves from one directional change of price to the next. We use the SMQ to quantify the FX market and evaluate the performance of the proposed methodology at major news announcem… ▽ More We define a methodology to quantify market activity on a 24 hour basis by defining a scale, the so-called scale of market quakes (SMQ). The SMQ is designed within a framework where we analyse the dynamics of excess price moves from one directional change of price to the next. We use the SMQ to quantify the FX market and evaluate the performance of the proposed methodology at major news announcements. The evolution of SMQ magnitudes from 2003 to 2009 is analysed across major currency pairs. △ Less

Submitted 9 September, 2009; originally announced September 2009.

Comments: 9 pages, 4 figures

Patterns in high-frequency FX data: Discovery of 12 empirical scaling laws

Authors: J. B. Glattfelder, A. Dupuis, R. B. Olsen

Abstract: We have discovered 12 independent new empirical scaling laws in foreign exchange data-series that hold for close to three orders of magnitude and across 13 currency exchange rates. Our statistical analysis crucially depends on an event-based approach that measures the relationship between different types of events. The scaling laws give an accurate estimation of the length of the price-curve coast… ▽ More We have discovered 12 independent new empirical scaling laws in foreign exchange data-series that hold for close to three orders of magnitude and across 13 currency exchange rates. Our statistical analysis crucially depends on an event-based approach that measures the relationship between different types of events. The scaling laws give an accurate estimation of the length of the price-curve coastline, which turns out to be surprisingly long. The new laws substantially extend the catalogue of stylised facts and sharply constrain the space of possible theoretical explanations of the market mechanisms. △ Less

Submitted 22 June, 2010; v1 submitted 5 September, 2008; originally announced September 2008.

Comments: 26 pages, 3 figures, 23 tables,2nd version (text made more concise and readable, algorithm pseudocode, results unchanged), 5-year datasets (USD-JPY, EUR-USD) provided at http://www.olsen.ch/more/datasets/

Journal ref: Quant. Financ. 11(4), 599 - 614 (2011) - published online Oct. 2010