Automatic mapping of free texts to bioinformatics ontology terms
Name
Erik Jaaniso
Abstract
In the field of bioinformatics, the number of tools and services is ever-increasing.
In order to make information about these resources available in a useful was, we annotate them with ontology terms.
This is currently done manually -- which is time-consuming and error-prone.
In this thesis, we set out to make a tool that helps the annotator by providing annotation suggestions.
We developed a program, that reads in free text descriptions of tools and services, adds content of web pages and publications related to the tool and based on this outputs best matches to ontology terms.
Then, we optimised the parameters of the program on manually done annotation sets.
Initial results look promising, as when comparing performance against these manual annotations, we see that many suggestions are agreeing with them.
Moreover, according to experienced annotators, many of the other suggestions are also correct.
In order to make information about these resources available in a useful was, we annotate them with ontology terms.
This is currently done manually -- which is time-consuming and error-prone.
In this thesis, we set out to make a tool that helps the annotator by providing annotation suggestions.
We developed a program, that reads in free text descriptions of tools and services, adds content of web pages and publications related to the tool and based on this outputs best matches to ontology terms.
Then, we optimised the parameters of the program on manually done annotation sets.
Initial results look promising, as when comparing performance against these manual annotations, we see that many suggestions are agreeing with them.
Moreover, according to experienced annotators, many of the other suggestions are also correct.
Graduation Thesis language
English
Graduation Thesis type
Master - Computer Science
Supervisor(s)
Hedi Peterson
Defence year
2016