dc.contributor.advisor |
Ranathunga L |
|
dc.contributor.author |
Priyadarshana YHPP |
|
dc.date.accessioned |
2022 |
|
dc.date.available |
2022 |
|
dc.date.issued |
2022 |
|
dc.identifier.citation |
Priyadarshana, Y.H.P.P. (2022). Named entity boundary detection for Sinhala [Master's theses, University of Moratuwa]. Institutional Repository University of Moratuwa. http://dl.lib.uom.lk/handle/123/22447 |
|
dc.identifier.uri |
http://dl.lib.uom.lk/handle/123/22447 |
|
dc.description.abstract |
Named entity recognition (NER) can be introduced as a sequential categorizing task
which contains a potential gravity in novel research arena. NER can be mentioned as
the foundation for accomplishing most common natural language processing (NLP)
tasks such as information extraction, information retrieval, semantic role labelling etc.
Even though plenty of attempts have been employed on NE type detection, still there
are plenty of avenues to be discovered under the NE boundary detection. Analyzing
Sinhala related contents which have been published in social media can also be
considered as one of the rising factors due to several human involvements in the recent
past. The ultimate goal which is to obtain a constructive deep neural framework for
determining named entity boundary detection has been achieved in a comprehensive
manner and the model has been tested using Sinhala related statements which have
been extracted through social media. Several objectives have been determined to
accomplish this task considering the existing baselines. Several novelties have been
identified to show off the uniqueness of this approach. Specifically, the novel concept
“Boundary Bubbles” has been used to identify the specific entity mentions considering
each head word for the identified named entities. Various experiments have been
conducted based on multiple evaluation criteria and the named entity boundary
detection model performs well with an average of 71% in Precision, 67% in Recall and
63% in F1 over the existing benchmarks. Hence this novel framework can be accepted
as a vital solution for determining named entity boundary detection under forecasting
various computational activities in social media. |
en_US |
dc.language.iso |
en |
en_US |
dc.subject |
DEEP NEURAL NETWORK |
en_US |
dc.subject |
NAMED ENTITIES |
en_US |
dc.subject |
NAMED ENTITY BOUNDARY |
en_US |
dc.subject |
NAMED ENTITY RECOGNITION |
en_US |
dc.subject |
NAMED ENTITY TYPE |
en_US |
dc.subject |
COMPUTER SCIENCE & ENGINEERING - Dissertation |
en_US |
dc.subject |
INFORMATION TECHNOLOGY - Dissertation |
en_US |
dc.subject |
COMPUTER SCIENCE- Dissertation |
en_US |
dc.title |
Named entity boundary detection for Sinhala |
en_US |
dc.type |
Thesis-Abstract |
en_US |
dc.identifier.faculty |
IT |
en_US |
dc.identifier.degree |
MSc in Information Technology By research |
en_US |
dc.identifier.department |
Department of Information Technology |
en_US |
dc.date.accept |
2022 |
|
dc.identifier.accno |
Th5080 |
en_US |