### Abstract

Knowledge based artificial neural networks offer an approach for connectionist theory refinement. We present an algorithm for refining and extending the domain theory incorporated in a knowledge based neural network using constructive neural network learning algorithms. The initial domain theory comprising of propositional rules is translated into a knowledge based network of threshold logic units (threshold neuron). The domain theory is modified by dynamically adding neurons to the existing network. A constructive neural network learning algorithm is used to add and train these additional neurons using a sequence of labeled examples. We propose a novel hybrid constructive learning algorithm based on the Tiling and Pyramid constructive learning algorithms that allows knowledge based neural network to handle patterns with continuous valued attributes. Results of experiments on two non-trivial tasks (the ribosome binding site prediction and the financial advisor) show that our algorithm compares favorably with other algorithms for connectionist theory refinement both in terms of generalization accuracy and network size.

Original language | English (US) |
---|---|

Title of host publication | IEEE World Congress on Computational Intelligence |

Editors | Anon |

Publisher | IEEE |

Pages | 2318-2323 |

Number of pages | 6 |

Volume | 3 |

State | Published - 1998 |

Event | Proceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3) - Anchorage, AK, USA Duration: May 4 1998 → May 9 1998 |

### Other

Other | Proceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3) |
---|---|

City | Anchorage, AK, USA |

Period | 5/4/98 → 5/9/98 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Software

### Cite this

*IEEE World Congress on Computational Intelligence*(Vol. 3, pp. 2318-2323). IEEE.

}

*IEEE World Congress on Computational Intelligence.*vol. 3, IEEE, pp. 2318-2323, Proceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3), Anchorage, AK, USA, 5/4/98.

**Constructive theory refinement in knowledge based neural networks.** / Parekh, Rajesh; Honavar, Vasant.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Constructive theory refinement in knowledge based neural networks

AU - Parekh, Rajesh

AU - Honavar, Vasant

PY - 1998

Y1 - 1998

N2 - Knowledge based artificial neural networks offer an approach for connectionist theory refinement. We present an algorithm for refining and extending the domain theory incorporated in a knowledge based neural network using constructive neural network learning algorithms. The initial domain theory comprising of propositional rules is translated into a knowledge based network of threshold logic units (threshold neuron). The domain theory is modified by dynamically adding neurons to the existing network. A constructive neural network learning algorithm is used to add and train these additional neurons using a sequence of labeled examples. We propose a novel hybrid constructive learning algorithm based on the Tiling and Pyramid constructive learning algorithms that allows knowledge based neural network to handle patterns with continuous valued attributes. Results of experiments on two non-trivial tasks (the ribosome binding site prediction and the financial advisor) show that our algorithm compares favorably with other algorithms for connectionist theory refinement both in terms of generalization accuracy and network size.

AB - Knowledge based artificial neural networks offer an approach for connectionist theory refinement. We present an algorithm for refining and extending the domain theory incorporated in a knowledge based neural network using constructive neural network learning algorithms. The initial domain theory comprising of propositional rules is translated into a knowledge based network of threshold logic units (threshold neuron). The domain theory is modified by dynamically adding neurons to the existing network. A constructive neural network learning algorithm is used to add and train these additional neurons using a sequence of labeled examples. We propose a novel hybrid constructive learning algorithm based on the Tiling and Pyramid constructive learning algorithms that allows knowledge based neural network to handle patterns with continuous valued attributes. Results of experiments on two non-trivial tasks (the ribosome binding site prediction and the financial advisor) show that our algorithm compares favorably with other algorithms for connectionist theory refinement both in terms of generalization accuracy and network size.

UR - http://www.scopus.com/inward/record.url?scp=0031630637&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031630637&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0031630637

VL - 3

SP - 2318

EP - 2323

BT - IEEE World Congress on Computational Intelligence

A2 - Anon, null

PB - IEEE

ER -