Knowledge-based TDNN architectures for features recognition in DNA sequences

Volume

4

fYear

2001

fDate

15-19 July 2001

Firstpage

2327

Abstract

We present work on the eflective utilization of ANNs for the recognition and location of specific regulatoly elements in genomic DNA, namely promoters. In particular, we rely on the utilization of knowledge-based NN (kbNN) techniques for the automatic construction of Time Delay NN (TDNN) architectures. The prior-knowledge is based and extracted from position weight-matrices of the DNA-sequence bases. The resulted kPRNN: knowledgeable Promoter Recognition Neural Network is trained, and the knowledge used to initialize the TDNN is refined. Experimental results show a significant improvement in TDNN-based promoter recognition.

Keywords

learning (artificial intelligence); neural nets; adaptive critics; aerospace applications; architectures; artificial systems; associative memory; automatic target recognition; biological foundations; biomedical applications; complexity analysis; computational intelligence; cortical networks; data compression; data mining; environmental data processing; evolutionary computations; evolutionary programming; fuzzy systems; global representations; handwriting recognition; image processing; industrial applications; information representation; intelligent agents; military applications; morphological neural networks; neural net competition; neural net projects; neurocoding; neurocontrol; neurodynamics; neurofuzzy systems; nonlinear analysis; nonparametric information theoretic learning algorithms; optimisation; precollege students; quantum neurocomputing; radial basis functions; reinforcement learning; robotics; service sector; signal processing; speech recognition; stability; supervised learning; support vector machines; system identification; target tracking; telecommunication; temporal processing; unsupervised learning; vision; Aerospace industry; Algorithm design and analysis; Artificial neural networks; Biological information theory; Image analysis; Information analysis; Neural networks; Signal analysis; Signal processing algorithms; Speech analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2001. Proceedings. IJCNN '01. International Joint Conference on

Conference_Location

Washington, DC, USA

ISSN

1098-7576

Print_ISBN

0-7803-7044-9

Type

conf

DOI

10.1109/IJCNN.2001.938723

Filename

938723