Probability distributions, trading strategies and leverage: an average of Guassian mixture models
Article Abstract:
The Guassian mixture model (a neural network to forecast density functions) is applied to a one-day-ahead forecasting task of the EUR/USD time series, benchmarking it against standard forecasting models like a naive model, a moving average convergence divergence technical model (MACD), an auto regressive moving average model (ARMA), a logistic regression model (LOGIT) and a multi layer network. The possibilities of improving the trading performance of these models with confirmation filters and leverage are examined. The Guassian mixture model outperforms all the benchmark models when taking advantage of the possibilities offered by a combination of more sophisticated trading strategies and leverage, mainly due to its ability to identify successfully trades with a high Sharpe ratio.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 2004
User Contributions:
Comment about this article or add new information about this topic:
A fractal forecasting model for financial time series
Article Abstract:
Financial market time series exhibit high degrees of non-linear variability and frequently have fractal properties. A multivariate system like financial markets, fractality is stochastic rather than deterministic and generally originates as a result of multiplicative interactions. The state transition -fitted residual scale ratio (ST-FRSR) reduces the predictive error primarily by capturing extreme fluctuations more accurately. The forecast error during the outlying events is reduced by between two-fifths and three-fifths, a highly significant outcome.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 2004
User Contributions:
Comment about this article or add new information about this topic:
Model uncertainty, thick modeling and the predictability of stock returns
Article Abstract:
Pesaran and Timmermann (P&T) consider various economic factors for forecasting stock returns. 'Recursive modeling' is applied, according to which the best forecasting model is chosen based upon the given statistical criterion. The method of selecting the best specification from all the available specifications is known as thin modeling. It focuses on modeling the decision in real time for excess returns.
Publication Name: Journal of Forecasting
Subject: Mathematics
ISSN: 0277-6693
Year: 2005
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: Profitability of price and quantity strategies in an oligopoly. Stochastically stable states in an oligopoly with differentiated goods: equivalence of price and quantity strategies
- Abstracts: Information theoretic approaches to inference in moment condition models. An efficient semiparametric estimator for binary response models
- Abstracts: Cost sharing: efficiency and implementation. An essay on the state of economic science. General equilibrium comparative statics: discrete shocks in production economies