LNCS 2810 Similarity Based Neural Networks for
Applications in Computational Molecular Biology
1ST EDITON BY Igor Fischer ISBN 9783540408130
pdf download
https://ebookball.com/product/lncs-2810-similarity-based-neural-
networks-for-applications-in-computational-molecular-biology-1st-
editon-by-igor-fischer-isbn-9783540408130-9096/
Explore and download more ebooks or textbooks
at ebookball.com
, Get Your Digital Files Instantly: PDF, ePub, MOBI and More
Quick Digital Downloads: PDF, ePub, MOBI and Other Formats
LNCS 2810 A Multiagent Based Constructive Approach for Feedforward
Neural Networks 1st Edition by Clodoaldo Lima, André Coelho, Fernando
Von Zuben ISBN 3540452311 9783540452317
https://ebookball.com/product/lncs-2810-a-multiagent-based-
constructive-approach-for-feedforward-neural-networks-1st-
edition-by-clodoaldo-lima-andra-c-coelho-fernando-von-zuben-
isbn-3540452311-9783540452317-11784/
LNCS 2810 Similarity Based Classification 1st edition by Axel Bernal,
Karen Hospevian, Tayfun Karadeniz, Jean Louis Lassez ISBN 3540408134
978-3540408130
https://ebookball.com/product/lncs-2810-similarity-based-
classification-1st-edition-by-axel-bernal-karen-hospevian-tayfun-
karadeniz-jean-louis-lassez-isbn-3540408134-978-3540408130-13434/
Laboratory Techniques in Biochemistry and Molecular Biology 1st editon
by Fallon, Booth, Bell ISBN 0444808620 9780444808622
https://ebookball.com/product/laboratory-techniques-in-
biochemistry-and-molecular-biology-1st-editon-by-fallon-booth-
bell-isbn-0444808620-9780444808622-8614/
Oral Biology Molecular Techniques and Applications Methods in
Molecular Biology 2010th Edition by Gregory Seymour, Mary Cullinan,
Nicholas Heng ISBN 1607618192 9781607618195
https://ebookball.com/product/oral-biology-molecular-techniques-
and-applications-methods-in-molecular-biology-2010th-edition-by-
gregory-seymour-mary-cullinan-nicholas-heng-
isbn-1607618192-9781607618195-7074/
,Artificial Neural Networks In Vehicular Pollution Modelling Studies in
Computational Intelligence 41 1st edition by Mukesh Khare, Shiva
Nagendra ISBN 3540374175 Â 978-3540374176
https://ebookball.com/product/artificial-neural-networks-in-
vehicular-pollution-modelling-studies-in-computational-
intelligence-41-1st-edition-by-mukesh-khare-shiva-nagendra-
isbn-3540374175-978-3540374176-19542/
(Ebook PDF) Computational Linguistics Models Resources Applications
1st edition by Igor Bolshakov, Alexander Gelbuk 9703601472 full
chapters
https://ebookball.com/product/ebook-pdf-computational-
linguistics-models-resources-applications-1st-edition-by-igor-
bolshakov-alexander-gelbuk-9703601472-full-chapters-8996/
(Ebook PDF) Computational Linguistics Models Resources Applications
1st edition by Igor Bolshakov, Alexander Gelbukh 9703601472 full
chapters
https://ebookball.com/product/ebook-pdf-computational-
linguistics-models-resources-applications-1st-edition-by-igor-
bolshakov-alexander-gelbukh-9703601472-full-chapters-14818/
LNCS 2834 Pattern Classification with Parallel Processing of the
Cellular Neural Networks Based Dynamic Programming 1st edition by
Hyongsuk Kim, Taewan Oh, Sangik Na, Changbae Yoon ISBN 3540200541
978-3540200543
https://ebookball.com/product/lncs-2834-pattern-classification-
with-parallel-processing-of-the-cellular-neural-networks-based-
dynamic-programming-1st-edition-by-hyongsuk-kim-taewan-oh-sangik-
na-changbae-yoon-isbn-3540200541-978-35/
Hierarchical Neural Networks for Image Interpretation 1st edition by
Sven Behnke 3540451692 9783540451693
https://ebookball.com/product/hierarchical-neural-networks-for-
image-interpretation-1st-edition-by-sven-
behnke-3540451692-9783540451693-19930/
, Similarity-Based Neural Networks for
Applications in Computational Molecular
Biology
Igor Fischer
Wilhelm-Schickard-Institut für Informatik, Universität Tübingen,
Sand 1, 72076 Tübingen, Germany
Abstract. This paper presents an alternative to distance-based neural
networks. A distance measure is the underlying property on which many
neural models rely, for example self-organizing maps or neural gas. How-
ever, a distance measure implies some requirements on the data which
are not always easy to satisfy in practice. This paper shows that a weaker
measure, the similarity measure, is sufficient in many cases. As an ex-
ample, similarity-based networks for strings are presented. Although a
metric can also be defined on strings, similarity is the established mea-
sure in string-intensive research, like computational molecular biology.
Similarity-based neural networks process data based on the same crite-
ria as other tools for analyzing DNA or amino-acid sequences.
1 Introduction
In respect to underlying mathematical properties, most artificial neural networks
used today can be classified as scalar product-based or distance-based. Of these,
multi-layer perceptrons and LVQ [1] are typical representatives.
In distance-based models, each neuron is assigned a pattern to which it is
sensitive. Appearance of the same or a similar pattern on the input results in
a high activation of that neuron — similarity being here understood as the
opposite of distance.
For numerical data, distance-based neural networks can easily be defined,
for there is a wide choice of distance measures, the Euclidean distance being
certainly the best known. In some applications, however, the data cannot be
represented as numbers or vectors. Although it may sometimes still be possible
to define a distance on them, such a measure is not always natural, in the sense
that it well represents relationships between data.
One such example are symbol strings, like DNA or amino-acid sequences
which are often subject to research in computational molecular biology. There, a
different measure – similarity – is usually used. It takes into account mutability of
symbols, which is determined through complex observations on many biologically
close sequences. To process such sequences with neural networks, it is preferable
to use a measure which is well empirically founded.
M.R. Berthold et al. (Eds.): IDA 2003, LNCS 2810, pp. 208–218, 2003.
c Springer-Verlag Berlin Heidelberg 2003
Applications in Computational Molecular Biology
1ST EDITON BY Igor Fischer ISBN 9783540408130
pdf download
https://ebookball.com/product/lncs-2810-similarity-based-neural-
networks-for-applications-in-computational-molecular-biology-1st-
editon-by-igor-fischer-isbn-9783540408130-9096/
Explore and download more ebooks or textbooks
at ebookball.com
, Get Your Digital Files Instantly: PDF, ePub, MOBI and More
Quick Digital Downloads: PDF, ePub, MOBI and Other Formats
LNCS 2810 A Multiagent Based Constructive Approach for Feedforward
Neural Networks 1st Edition by Clodoaldo Lima, André Coelho, Fernando
Von Zuben ISBN 3540452311 9783540452317
https://ebookball.com/product/lncs-2810-a-multiagent-based-
constructive-approach-for-feedforward-neural-networks-1st-
edition-by-clodoaldo-lima-andra-c-coelho-fernando-von-zuben-
isbn-3540452311-9783540452317-11784/
LNCS 2810 Similarity Based Classification 1st edition by Axel Bernal,
Karen Hospevian, Tayfun Karadeniz, Jean Louis Lassez ISBN 3540408134
978-3540408130
https://ebookball.com/product/lncs-2810-similarity-based-
classification-1st-edition-by-axel-bernal-karen-hospevian-tayfun-
karadeniz-jean-louis-lassez-isbn-3540408134-978-3540408130-13434/
Laboratory Techniques in Biochemistry and Molecular Biology 1st editon
by Fallon, Booth, Bell ISBN 0444808620 9780444808622
https://ebookball.com/product/laboratory-techniques-in-
biochemistry-and-molecular-biology-1st-editon-by-fallon-booth-
bell-isbn-0444808620-9780444808622-8614/
Oral Biology Molecular Techniques and Applications Methods in
Molecular Biology 2010th Edition by Gregory Seymour, Mary Cullinan,
Nicholas Heng ISBN 1607618192 9781607618195
https://ebookball.com/product/oral-biology-molecular-techniques-
and-applications-methods-in-molecular-biology-2010th-edition-by-
gregory-seymour-mary-cullinan-nicholas-heng-
isbn-1607618192-9781607618195-7074/
,Artificial Neural Networks In Vehicular Pollution Modelling Studies in
Computational Intelligence 41 1st edition by Mukesh Khare, Shiva
Nagendra ISBN 3540374175 Â 978-3540374176
https://ebookball.com/product/artificial-neural-networks-in-
vehicular-pollution-modelling-studies-in-computational-
intelligence-41-1st-edition-by-mukesh-khare-shiva-nagendra-
isbn-3540374175-978-3540374176-19542/
(Ebook PDF) Computational Linguistics Models Resources Applications
1st edition by Igor Bolshakov, Alexander Gelbuk 9703601472 full
chapters
https://ebookball.com/product/ebook-pdf-computational-
linguistics-models-resources-applications-1st-edition-by-igor-
bolshakov-alexander-gelbuk-9703601472-full-chapters-8996/
(Ebook PDF) Computational Linguistics Models Resources Applications
1st edition by Igor Bolshakov, Alexander Gelbukh 9703601472 full
chapters
https://ebookball.com/product/ebook-pdf-computational-
linguistics-models-resources-applications-1st-edition-by-igor-
bolshakov-alexander-gelbukh-9703601472-full-chapters-14818/
LNCS 2834 Pattern Classification with Parallel Processing of the
Cellular Neural Networks Based Dynamic Programming 1st edition by
Hyongsuk Kim, Taewan Oh, Sangik Na, Changbae Yoon ISBN 3540200541
978-3540200543
https://ebookball.com/product/lncs-2834-pattern-classification-
with-parallel-processing-of-the-cellular-neural-networks-based-
dynamic-programming-1st-edition-by-hyongsuk-kim-taewan-oh-sangik-
na-changbae-yoon-isbn-3540200541-978-35/
Hierarchical Neural Networks for Image Interpretation 1st edition by
Sven Behnke 3540451692 9783540451693
https://ebookball.com/product/hierarchical-neural-networks-for-
image-interpretation-1st-edition-by-sven-
behnke-3540451692-9783540451693-19930/
, Similarity-Based Neural Networks for
Applications in Computational Molecular
Biology
Igor Fischer
Wilhelm-Schickard-Institut für Informatik, Universität Tübingen,
Sand 1, 72076 Tübingen, Germany
Abstract. This paper presents an alternative to distance-based neural
networks. A distance measure is the underlying property on which many
neural models rely, for example self-organizing maps or neural gas. How-
ever, a distance measure implies some requirements on the data which
are not always easy to satisfy in practice. This paper shows that a weaker
measure, the similarity measure, is sufficient in many cases. As an ex-
ample, similarity-based networks for strings are presented. Although a
metric can also be defined on strings, similarity is the established mea-
sure in string-intensive research, like computational molecular biology.
Similarity-based neural networks process data based on the same crite-
ria as other tools for analyzing DNA or amino-acid sequences.
1 Introduction
In respect to underlying mathematical properties, most artificial neural networks
used today can be classified as scalar product-based or distance-based. Of these,
multi-layer perceptrons and LVQ [1] are typical representatives.
In distance-based models, each neuron is assigned a pattern to which it is
sensitive. Appearance of the same or a similar pattern on the input results in
a high activation of that neuron — similarity being here understood as the
opposite of distance.
For numerical data, distance-based neural networks can easily be defined,
for there is a wide choice of distance measures, the Euclidean distance being
certainly the best known. In some applications, however, the data cannot be
represented as numbers or vectors. Although it may sometimes still be possible
to define a distance on them, such a measure is not always natural, in the sense
that it well represents relationships between data.
One such example are symbol strings, like DNA or amino-acid sequences
which are often subject to research in computational molecular biology. There, a
different measure – similarity – is usually used. It takes into account mutability of
symbols, which is determined through complex observations on many biologically
close sequences. To process such sequences with neural networks, it is preferable
to use a measure which is well empirically founded.
M.R. Berthold et al. (Eds.): IDA 2003, LNCS 2810, pp. 208–218, 2003.
c Springer-Verlag Berlin Heidelberg 2003
