Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction
The development of Web 2.0 has improved peoples’ ability to share their sentiments, or opinions, on various services or products easily. This is to investigate the public opinions that are expressed within the reviews. Aspect-based sentiment analysis (ABSA) deemed to receive a set of texts (e.g., pr...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | http://eprints.usm.my/52458/1/AL%20JANABI%20OMAR%20MUSTAFA%20ABBAS%20-%20TESIS24.pdf |
| _version_ | 1797012820767801344 |
|---|---|
| author | Abbas, Al Janabi Omar Mustafa |
| author_facet | Abbas, Al Janabi Omar Mustafa |
| author_sort | Abbas, Al Janabi Omar Mustafa |
| collection | USM |
| description | The development of Web 2.0 has improved peoples’ ability to share their sentiments, or opinions, on various services or products easily. This is to investigate the public opinions that are expressed within the reviews. Aspect-based sentiment analysis (ABSA) deemed to receive a set of texts (e.g., product reviews or online reviews) and identify the opinion-target (aspect) within each review. Contemporary aspect-based sentiment analysis systems, like the aspect grouping, rely predominantly on lexicon-based and manually labelled seeds that is being incorporated into the topic models. The previously developed systems for Aspect Category Detection (ACD) rely mostly on supervised machine learning techniques. The problem of implicit aspect extraction is being addressed using either pre-constructed rules or pre-labelled clues for performing implicit aspect detection. To cope with these issues, Bayesian probabilistic models proposed to perform the aspect grouping, ACD, and distributed vectors for implicit aspect extraction. Parametric and non-parametric Bayesian models are developed to conduct both the annotated and non-annotated data, that are; Topic-seeds Latent Dirichlet allocation (TSLDA) and Hierarchical Dirichlet Process-Collapsed Gibbs Sampling (HDP-CGS), respectively. The yielded aspect groups using the developed Bayesian models fed into the advised distributed vector (i.e., Skip-gram) for implicit aspect extraction. The proposed methodology evaluated using several online reviews benchmark datasets (including datasets annotated using reviews retrieved from Amazon.com and TripAdvisor.com). |
| first_indexed | 2024-03-06T15:53:18Z |
| format | Thesis |
| id | usm.eprints-52458 |
| institution | Universiti Sains Malaysia |
| language | English |
| last_indexed | 2024-03-06T15:53:18Z |
| publishDate | 2021 |
| record_format | dspace |
| spelling | usm.eprints-524582022-04-30T15:40:06Z http://eprints.usm.my/52458/ Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction Abbas, Al Janabi Omar Mustafa QA75-76.95 Calculating Machines The development of Web 2.0 has improved peoples’ ability to share their sentiments, or opinions, on various services or products easily. This is to investigate the public opinions that are expressed within the reviews. Aspect-based sentiment analysis (ABSA) deemed to receive a set of texts (e.g., product reviews or online reviews) and identify the opinion-target (aspect) within each review. Contemporary aspect-based sentiment analysis systems, like the aspect grouping, rely predominantly on lexicon-based and manually labelled seeds that is being incorporated into the topic models. The previously developed systems for Aspect Category Detection (ACD) rely mostly on supervised machine learning techniques. The problem of implicit aspect extraction is being addressed using either pre-constructed rules or pre-labelled clues for performing implicit aspect detection. To cope with these issues, Bayesian probabilistic models proposed to perform the aspect grouping, ACD, and distributed vectors for implicit aspect extraction. Parametric and non-parametric Bayesian models are developed to conduct both the annotated and non-annotated data, that are; Topic-seeds Latent Dirichlet allocation (TSLDA) and Hierarchical Dirichlet Process-Collapsed Gibbs Sampling (HDP-CGS), respectively. The yielded aspect groups using the developed Bayesian models fed into the advised distributed vector (i.e., Skip-gram) for implicit aspect extraction. The proposed methodology evaluated using several online reviews benchmark datasets (including datasets annotated using reviews retrieved from Amazon.com and TripAdvisor.com). 2021-10 Thesis NonPeerReviewed application/pdf en http://eprints.usm.my/52458/1/AL%20JANABI%20OMAR%20MUSTAFA%20ABBAS%20-%20TESIS24.pdf Abbas, Al Janabi Omar Mustafa (2021) Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction. PhD thesis, Universiti Sains Malaysia. |
| spellingShingle | QA75-76.95 Calculating Machines Abbas, Al Janabi Omar Mustafa Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction |
| title | Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction |
| title_full | Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction |
| title_fullStr | Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction |
| title_full_unstemmed | Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction |
| title_short | Domain-independent Bayesian Model For Aspect Category Detection And Distributed Vector For Implicit Aspect Extraction |
| title_sort | domain independent bayesian model for aspect category detection and distributed vector for implicit aspect extraction |
| topic | QA75-76.95 Calculating Machines |
| url | http://eprints.usm.my/52458/1/AL%20JANABI%20OMAR%20MUSTAFA%20ABBAS%20-%20TESIS24.pdf |
| work_keys_str_mv | AT abbasaljanabiomarmustafa domainindependentbayesianmodelforaspectcategorydetectionanddistributedvectorforimplicitaspectextraction |