Modeling the Syntax of the song of the Great Reed Warbler Faculty of Engineering, LTH
(2014) FMSL01 20142Mathematical Statistics
- Abstract (Swedish)
- The song of many songbirds can be thought of as consisting of variable sequences of a finite set of syllables.
A common approach in understanding the structure of these songs is to model the syllable sequences
with a Markov Model. The Markov Model can either allow one-to-one (Markov Chain), many-to-many
(Hidden Markov Model) or many-to-one (Partially Observed Markov Model) state to syllable mappings.
In this project the song of the Great Reed Warbler is being studied in terms of the syllable sequences
(strophes) being generated. It is shown that the Markov chain captures a lot of the structure in the song
in the sense that it to large degree reproduces the syllable distributions at a specific position in the song
that were observed... (More) - The song of many songbirds can be thought of as consisting of variable sequences of a finite set of syllables.
A common approach in understanding the structure of these songs is to model the syllable sequences
with a Markov Model. The Markov Model can either allow one-to-one (Markov Chain), many-to-many
(Hidden Markov Model) or many-to-one (Partially Observed Markov Model) state to syllable mappings.
In this project the song of the Great Reed Warbler is being studied in terms of the syllable sequences
(strophes) being generated. It is shown that the Markov chain captures a lot of the structure in the song
in the sense that it to large degree reproduces the syllable distributions at a specific position in the song
that were observed in data. The repetition distribution for some syllable classes was consistent with
that of a Markov chain while other syllable classes were better modeled by allowing the self-transition
probability to be adapted as the syllable class is repeated more and more. Still some other syllable
classes did not have their repetition distributions accurately captured by these two alternatives. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/4647240
- author
- Tronarp, Filip
- supervisor
- organization
- course
- FMSL01 20142
- year
- 2014
- type
- M2 - Bachelor Degree
- subject
- language
- English
- id
- 4647240
- date added to LUP
- 2014-09-16 11:37:35
- date last changed
- 2014-09-16 11:37:35
@misc{4647240, abstract = {{The song of many songbirds can be thought of as consisting of variable sequences of a finite set of syllables. A common approach in understanding the structure of these songs is to model the syllable sequences with a Markov Model. The Markov Model can either allow one-to-one (Markov Chain), many-to-many (Hidden Markov Model) or many-to-one (Partially Observed Markov Model) state to syllable mappings. In this project the song of the Great Reed Warbler is being studied in terms of the syllable sequences (strophes) being generated. It is shown that the Markov chain captures a lot of the structure in the song in the sense that it to large degree reproduces the syllable distributions at a specific position in the song that were observed in data. The repetition distribution for some syllable classes was consistent with that of a Markov chain while other syllable classes were better modeled by allowing the self-transition probability to be adapted as the syllable class is repeated more and more. Still some other syllable classes did not have their repetition distributions accurately captured by these two alternatives.}}, author = {{Tronarp, Filip}}, language = {{eng}}, note = {{Student Paper}}, title = {{Modeling the Syntax of the song of the Great Reed Warbler Faculty of Engineering, LTH}}, year = {{2014}}, }