Subjectivity detection in sentiment analysis for Arabic educational domain based on feature and lexicon-based approach

Aqdas Enad Eesee (P72221)

Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/476294

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Nazlia Omar, Prof. Dr.
dc.contributor.author	Aqdas Enad Eesee (P72221)
dc.date.accessioned	2023-10-06T09:15:58Z	-
dc.date.available	2023-10-06T09:15:58Z	-
dc.date.issued	2016-02-25
dc.identifier.other	ukmvital:81392
dc.identifier.uri	https://ptsldigital.ukm.my/jspui/handle/123456789/476294	-
dc.description	Sentiment analysis is the process of identifying opinions of consumers toward particular product or service in order to enhance the quality of such product or service. Recently, education domain has caught many researchers’ attentions by classifying students’ reviews in order to enhance the process of learning. Since the reviews are being provided by students thus, several objective or so-called factual sentences could be existed. These sentences do not indicate an opinion which leads them to be incorrectly classified. Therefore, it is necessary to accommodate subjectivity detection approach in order to eliminate the objective sentences. Several approaches have been proposed for subjectivity detection. Yet, there is still room for improvement in terms of utilizing appropriate features that have the ability to distinguish subjective and objective sentences. The dataset used in this study is an academic dataset that contains students’ reviews from many colleges. The proposed features that are used for subjectivity detection consist of POS tagging, keywords, negation and entity features. Furthermore, an Arabic lexicon has been used in order to provide the polarity for each review. Support Vector Machine classifier has been used to classify the review into positive and negative classes. For this manner, unigram and bigram approaches have been utilized as a feature space for the SVM. The evaluation has been performed based on precision, recall and f-measure in which a comparison is established between the use of SVM with the proposed subjectivity detection and without the proposed method. Results show that the use of SVM with the proposed subjectivity detection method has outperformed the use of SVM without the proposed method by achieving 84% of an overall f-measure. Such results demonstrate the usability of the proposed subjectivity detection.,Master of Computer Science
dc.language.iso	eng
dc.publisher	UKM, Bangi
dc.relation	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat
dc.rights	UKM
dc.subject	Sentiment analysis
dc.subject	Subjectivity detection
dc.subject	Education domain
dc.subject	Dissertations, Academic -- Malaysia
dc.title	Subjectivity detection in sentiment analysis for Arabic educational domain based on feature and lexicon-based approach
dc.type	theses
dc.format.pages	114
dc.identifier.barcode	002186(2016)
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Files in This Item:

File	Description	Size	Format
ukmvital_81392+SOURCE1+SOURCE1.0.PDF Restricted Access		127.1 kB	Adobe PDF	View/Open

Show simple item record Recommend this item