Nonlinear Modelling of High Frequency Financial Time Series Edited by Christian Dunis and Bin Zhou In the competitive and risky environment of today's financial markets, daily prices and models based upon low frequency price series data do not provide the level of accuracy required by traders and a growing number of risk managers. To improve results, more and more researchers and practitioners are turning to high frequency data. Nonlinear Modelling of High Frequency Financial Time Series presents the latest developments and views of leading international researchers and market practitioners, in modelling high frequency data in finance. Combining both nonlinear modelling and intraday data for financial markets, the editors provide a fascinating foray into this extremely popular discipline. This book evolves around four major themes. The first introductory section focuses on high frequency financial data. The second part examines the exact nature of the time series considered: several linearity tests are presented and applied and their modelling implications assessed. The third and fourth parts are dedicated to modelling and forecasting these financial time series.

"Dunis, Timmermann and Moody have assembled a fascinating and insightful collection of articles that address the latest developments in forecast combination, long memory, and portfolio choice. These papers present state-of-the-art estimation techniques along with relevant empirical applications. Both academics and practitioners will find this anthology to be a worthy addition to their libraries." — Joshua Rosenberg, NYU - Stern School of Business

"A modern book on financial econometrics has to consider interesting and relevant topics from the viewpoint of recently developed techniques that have been shown to actually work. This book delivers in all aspects&The many new techniques displayed include the use of loss functions based on economic rather than statistical criteria, the benefits of combining forecasts, dimension reduction, structural changes, and long memory fractional cointegration, neural networks, and high frequency data techniques applied to exchange rates. The authors provide plenty to think about." — Professor Clive W J Granger, University of California, San Diego

"This book is a welcome addition to the literature on forecasting in financial markets." — Professor Ken Holden, Liverpool Business School

Developments in Forecast Combination and Portfolio Choice brings together papers that address current frontier research within the field of quantitative finance. Focusing on three core themes of model and forecast combinations; structural change and long memory; and controlling downside risk and investment strategies, the book provides an authoritative collection of readings that are relevant to academics and practitioners alike.

Today’ s financial markets are characterised by a large number of participants, with different appetites for risk, different time horizons, different motivations and reactions to unexpected news. The mathematical techniques and models used in the forecasting of financial markets have therefore grown ever more sophisticated as traders, analysts and investors seek to gain an edge on their competitors. Written by leading international researchers and practitioners, this book focuses on three major themes of today’ s state of the art financial research: modelling with high frequency data, the information content of volatility markets, and applications of neural networks and genetic algorithms to financial time series. Forecasting Financial Markets includes empirical applications to present the very latest thinking on these complex techniques, including:

• High frequency exchange rates

• Intraday volatility

• Autocorrelation and variance ratio tests

• Conditional volatility

• GARCH processes

• Chaotic systems

• Nonlinearity

• Stochastic and EXPAR models

• Artificial neural networks

• Genetic algorithms