## Predicting the exchange traded fund DIA with a combination of genetic algorithms and neural networks (2004)

Venue: | EXPERT SYSTEMS WITH APPLICATIONS |

Citations: | 8 - 1 self |

### BibTeX

@MISC{Versace04predictingthe,

author = {Massimiliano Versace and Rushi Bhatt and Oliver Hinds and Mark Shiffer},

title = {Predicting the exchange traded fund DIA with a combination of genetic algorithms and neural networks},

year = {2004}

}

### OpenURL

### Abstract

### Citations

7727 |
Genetic Algorithms
- Goldberg
- 1989
(Show Context)
Citation Context ...space of system configuration is provided by GAs. GAs are a class of probabilistic search techniques that has been developed in the past decades as a general-purpose optimization tool (Holland, 1975; =-=Goldberg, 1989-=-). GAs mimic biological evolution by using a massively parallel search mechanism that involves (a) the initialization of a random population of systems (or candidate solutions) (b) ordering the system... |

3885 |
Neural Networks: A Comprehensive Foundation
- Haykin
(Show Context)
Citation Context ... perspective of statistical learning theory, and they usually yield good performance when used for real-world data analysis or in predicting nonlinear dynamical systems (Lapedes & Farber, 1987, 1988; =-=Haykin, 1999-=-; Duda, Hart, & Stork, 2000). ANNs have good generalization capabilities and are usually robust against noisy or missing data, all of which are highly desirable properties time series prediction. Impo... |

862 |
Adaptation in Natural and Artificial Systems. Ann Arbor
- Holland
- 1975
(Show Context)
Citation Context ...his very large space of system configuration is provided by GAs. GAs are a class of probabilistic search techniques that has been developed in the past decades as a general-purpose optimization tool (=-=Holland, 1975-=-; Goldberg, 1989). GAs mimic biological evolution by using a massively parallel search mechanism that involves (a) the initialization of a random population of systems (or candidate solutions) (b) ord... |

747 | Hierarchical mixtures of experts and the EM algorithm
- Jordan, Jacobs
- 1994
(Show Context)
Citation Context ... classification methods, although it is difficult to say in general in what kind of tasks mixtures-of-experts or any other kind of model will perform well on any given dataset (Kang & Oh, 1997, 2000; =-=Jordan & Jacobs, 1994-=-; Jordan & Xu, 1995; Waterhouse, 2002). In the ANNs literature, inappropriate topology selection and weight training are frequently blamed for poor performance. Increasing the number of hidden layer n... |

325 | Distributed Representations, Simple Recurrent Networks
- Elman
- 1991
(Show Context)
Citation Context ...as well as the type of network and the input data type (database A or B; Table 3 for details). Every ith MON is composed of j networks to be selected from one of two types: Recurrent Backpropagation (=-=Elman, 1991-=-) or RBF networks (Duda et al., 2000). Elman networks are particularly suited for data when the temporal order of the data plays a crucial role (Elman, 1991), as it does in financial time series. RBF ... |

224 |
Designing Neural Networks Using Genetic Algorithms
- Miller, Todd, et al.
- 1989
(Show Context)
Citation Context ...en obtained in combining GAs and ANNs in hybrid systems (Belew, McInnernay, & Schraudolph, 1990; Montana & Davis, 1989; Shaffer, Whitely, & Eshelman, 1990; Chang & Lippmann, 1991; Harp & Samad,s1991; =-=Miller, Todd, & Hedge, 1989-=-; Whitley, 1989; Kingdon, 1997; Venkatesan & Kumar, 2002). This paper presents an application of such a hybrid system that uses GAs for selecting an appropriate combination of networks, parameters and... |

203 |
Training feed forward Neural Networks using Genetic Algorithms
- Montana, Davis
- 1989
(Show Context)
Citation Context ...for solving the problem if discovering appropriate parameterizations for ANNs, and good results have been obtained in combining GAs and ANNs in hybrid systems (Belew, McInnernay, & Schraudolph, 1990; =-=Montana & Davis, 1989-=-; Shaffer, Whitely, & Eshelman, 1990; Chang & Lippmann, 1991; Harp & Samad,s1991; Miller, Todd, & Hedge, 1989; Whitley, 1989; Kingdon, 1997; Venkatesan & Kumar, 2002). This paper presents an applicati... |

174 |
Nonlinear Signal Processing Using Neural Networks
- Lapedes, Farber
- 1987
(Show Context)
Citation Context ...und theoretical basis from the perspective of statistical learning theory, and they usually yield good performance when used for real-world data analysis or in predicting nonlinear dynamical systems (=-=Lapedes & Farber, 1987-=-, 1988; Haykin, 1999; Duda, Hart, & Stork, 2000). ANNs have good generalization capabilities and are usually robust against noisy or missing data, all of which are highly desirable properties time ser... |

142 | Generalization by weight elimination with application to forecasting - Weigend, Rumelhart, et al. - 1991 |

133 |
Threshold Autoregression, Limit Cycle and Cyclical Data (with discussion
- Tong, Lim
(Show Context)
Citation Context ... price trend of a given security. Parallel to the development if AI technologies, the last decade has also witnessed the development of several nonlinear time series models (Granger & Anderson, 1978, =-=Tong & Lim, 1980-=-, Engle, 1982). These nonlinear models, although not suffering from a priori assumption of a linear relationship between data and time series, are limited by the fact that an a priori assumption of th... |

101 |
Managerial application of neural networks: the case of bank failure prediction
- Tam, Kiang
- 1992
(Show Context)
Citation Context ...e on financial applications of ANNs (Trippi & Turban, 1993; Azoff, 1994; Refenes, 1995; Gately, 1996; Odom & Sharda, 1990; Coleman, Graettinger, & Lawrence 1991; Salchenkerger, Vinar, & Lash, 1992;s2 =-=Tam & Kiang, 1992-=-; Wilson & Sharda, 1994, Weigend, Rumelhart, & Hubermann, 1992; Zhang, 1998, for a review). Although encouraging results have been reported in which ANNs-based systems outperformed widely-used well-es... |

97 | Convergence results for the EM approach to mixtures of expert architectures. Neural Networks
- Jordan, Xu
- 1995
(Show Context)
Citation Context ..., although it is difficult to say in general in what kind of tasks mixtures-of-experts or any other kind of model will perform well on any given dataset (Kang & Oh, 1997, 2000; Jordan & Jacobs, 1994; =-=Jordan & Xu, 1995-=-; Waterhouse, 2002). In the ANNs literature, inappropriate topology selection and weight training are frequently blamed for poor performance. Increasing the number of hidden layer neurons helps improv... |

86 |
Neural Network Time Series Forecasting of Financial Market
- Azoff
- 1994
(Show Context)
Citation Context ...e the capability of discovering associations between features that may not have been expected or looked for. There is an extensive literature on financial applications of ANNs (Trippi & Turban, 1993; =-=Azoff, 1994-=-; Refenes, 1995; Gately, 1996; Odom & Sharda, 1990; Coleman, Graettinger, & Lawrence 1991; Salchenkerger, Vinar, & Lash, 1992;s2 Tam & Kiang, 1992; Wilson & Sharda, 1994, Weigend, Rumelhart, & Huberma... |

73 |
Introduction to bilinear time series models. Vandenhoeck and Ruprecht
- Granger, Anderson
- 1978
(Show Context)
Citation Context ... in the apparently chaotic price trend of a given security. Parallel to the development if AI technologies, the last decade has also witnessed the development of several nonlinear time series models (=-=Granger & Anderson, 1978-=-, Tong & Lim, 1980, Engle, 1982). These nonlinear models, although not suffering from a priori assumption of a linear relationship between data and time series, are limited by the fact that an a prior... |

72 | How neural nets work - Lapedes, Farber |

63 |
Bankruptcy prediction using neural networks
- Wilson, Sharda
- 1994
(Show Context)
Citation Context ...ications of ANNs (Trippi & Turban, 1993; Azoff, 1994; Refenes, 1995; Gately, 1996; Odom & Sharda, 1990; Coleman, Graettinger, & Lawrence 1991; Salchenkerger, Vinar, & Lash, 1992;s2 Tam & Kiang, 1992; =-=Wilson & Sharda, 1994-=-, Weigend, Rumelhart, & Hubermann, 1992; Zhang, 1998, for a review). Although encouraging results have been reported in which ANNs-based systems outperformed widely-used well-established statistical m... |

48 |
A neural network model for bankruptcy prediction
- Odom, R
- 1990
(Show Context)
Citation Context ...ns between features that may not have been expected or looked for. There is an extensive literature on financial applications of ANNs (Trippi & Turban, 1993; Azoff, 1994; Refenes, 1995; Gately, 1996; =-=Odom & Sharda, 1990-=-; Coleman, Graettinger, & Lawrence 1991; Salchenkerger, Vinar, & Lash, 1992;s2 Tam & Kiang, 1992; Wilson & Sharda, 1994, Weigend, Rumelhart, & Hubermann, 1992; Zhang, 1998, for a review). Although enc... |

42 | Neural networks: A new tool for predicting thrift failures - Salchenberger, Cinar, et al. - 1992 |

35 | Classification and Regression using Mixtures of Experts
- Waterhouse
- 1997
(Show Context)
Citation Context ...fficult to say in general in what kind of tasks mixtures-of-experts or any other kind of model will perform well on any given dataset (Kang & Oh, 1997, 2000; Jordan & Jacobs, 1994; Jordan & Xu, 1995; =-=Waterhouse, 2002-=-). In the ANNs literature, inappropriate topology selection and weight training are frequently blamed for poor performance. Increasing the number of hidden layer neurons helps improving network perfor... |

32 | Genetic synthesis of neural network architecture - Harp, Samad - 1991 |

26 |
Using genetic algorithms to improve pattern classification performance
- chang, Lippman
- 1991
(Show Context)
Citation Context ...rizations for ANNs, and good results have been obtained in combining GAs and ANNs in hybrid systems (Belew, McInnernay, & Schraudolph, 1990; Montana & Davis, 1989; Shaffer, Whitely, & Eshelman, 1990; =-=Chang & Lippmann, 1991-=-; Harp & Samad,s1991; Miller, Todd, & Hedge, 1989; Whitley, 1989; Kingdon, 1997; Venkatesan & Kumar, 2002). This paper presents an application of such a hybrid system that uses GAs for selecting an ap... |

20 |
Intelligent Systems and Financial Forecasting
- Kingdon
- 1997
(Show Context)
Citation Context ...rid systems (Belew, McInnernay, & Schraudolph, 1990; Montana & Davis, 1989; Shaffer, Whitely, & Eshelman, 1990; Chang & Lippmann, 1991; Harp & Samad,s1991; Miller, Todd, & Hedge, 1989; Whitley, 1989; =-=Kingdon, 1997-=-; Venkatesan & Kumar, 2002). This paper presents an application of such a hybrid system that uses GAs for selecting an appropriate combination of networks, parameters and training regimen for predicti... |

18 |
Neural Networks in Finance and Investment: Using Artificial Intelligence to Improve Real-world Performance
- Trippi, Turban
- 1993
(Show Context)
Citation Context ...sed neural networks have the capability of discovering associations between features that may not have been expected or looked for. There is an extensive literature on financial applications of ANNs (=-=Trippi & Turban, 1993-=-; Azoff, 1994; Refenes, 1995; Gately, 1996; Odom & Sharda, 1990; Coleman, Graettinger, & Lawrence 1991; Salchenkerger, Vinar, & Lash, 1992;s2 Tam & Kiang, 1992; Wilson & Sharda, 1994, Weigend, Rumelha... |

14 |
Neural network system for forecasting method selection. Decision Support Systems
- Chu, Widjaja
- 1994
(Show Context)
Citation Context ...Combining classifiers and boosting methods often lead to improvement in performance over single neural networks. Several studies (Pelikan, de Groot, & Wurtz, 1992; Ginzburg & Horn, 1994; Zhang, 1994; =-=Chu & Widjaja, 1994-=-) have shown improvement of performance in combining different ANNs or training similar ANNs on different features of the input data and then recombining their output in a later stage. Our usage of a ... |

9 | Neural networks for bankruptcy prediction: The power to solve financial problems - Coleman, Graettinger, et al. - 1991 |

9 | Power consumption in West-Bohemia: Improved forecasts with decorrelating connectionist networks - Groot, C, et al. - 1992 |

8 |
Combined neural networks for time series analysis
- Ginzburg, Horn
- 1994
(Show Context)
Citation Context ...bove chance in a given testing set. Combining classifiers and boosting methods often lead to improvement in performance over single neural networks. Several studies (Pelikan, de Groot, & Wurtz, 1992; =-=Ginzburg & Horn, 1994-=-; Zhang, 1994; Chu & Widjaja, 1994) have shown improvement of performance in combining different ANNs or training similar ANNs on different features of the input data and then recombining their output... |

8 | Statistical mechanics of the mixture of experts
- Kang, Oh
- 1997
(Show Context)
Citation Context ...al techniques and other classification methods, although it is difficult to say in general in what kind of tasks mixtures-of-experts or any other kind of model will perform well on any given dataset (=-=Kang & Oh, 1997-=-, 2000; Jordan & Jacobs, 1994; Jordan & Xu, 1995; Waterhouse, 2002). In the ANNs literature, inappropriate topology selection and weight training are frequently blamed for poor performance. Increasing... |

6 |
A genetic algorithms approach to growth phase forecasting of wireless subscribers
- Venkatesan, Kumar
- 2002
(Show Context)
Citation Context ...lew, McInnernay, & Schraudolph, 1990; Montana & Davis, 1989; Shaffer, Whitely, & Eshelman, 1990; Chang & Lippmann, 1991; Harp & Samad,s1991; Miller, Todd, & Hedge, 1989; Whitley, 1989; Kingdon, 1997; =-=Venkatesan & Kumar, 2002-=-). This paper presents an application of such a hybrid system that uses GAs for selecting an appropriate combination of networks, parameters and training regimen for predicting the direction of variat... |

4 | Fuzzyart: an adaptative resonance algorithm for rapid, stable classification of analog patterns - Carpenter, Grossberg, et al. - 1991 |

4 |
Autoregressive conditional to select inputs with estimates of the variance of UK inflation
- Engle
- 1982
(Show Context)
Citation Context ...given security. Parallel to the development if AI technologies, the last decade has also witnessed the development of several nonlinear time series models (Granger & Anderson, 1978, Tong & Lim, 1980, =-=Engle, 1982-=-). These nonlinear models, although not suffering from a priori assumption of a linear relationship between data and time series, are limited by the fact that an a priori assumption of the nature of t... |

4 |
On crossover os an evolutionary viable strategy
- Shaffer, Whitely, et al.
- 1990
(Show Context)
Citation Context ... if discovering appropriate parameterizations for ANNs, and good results have been obtained in combining GAs and ANNs in hybrid systems (Belew, McInnernay, & Schraudolph, 1990; Montana & Davis, 1989; =-=Shaffer, Whitely, & Eshelman, 1990-=-; Chang & Lippmann, 1991; Harp & Samad,s1991; Miller, Todd, & Hedge, 1989; Whitley, 1989; Kingdon, 1997; Venkatesan & Kumar, 2002). This paper presents an application of such a hybrid system that uses... |

4 | The GENITOR algorithms and selection pressure: why rank-based allocation of reproductive trials is the best - Whiteley - 1989 |

3 | Evolving networks: using GAs with connectioninst learning - Below, McInnernay, et al. - 1990 |

1 | Versace et al. / Expert Systems with Applications xx (2004) xxx–xxx 9 Refenes - M - 1995 |