New Approach of Sarcasm Detection In Indonesian Marketplace Product Review

Product reviews on e-commerce become very important for prospective buyers to determine the items to be bought. Unfortunately, the results of the product reviews are not accurate because of the meaningful reviews of sarcasm. Sarcasm is a form of intelligent speech where the speaker or writer says or...

Full description

Saved in:
Bibliographic Details
Published in2020 6th International Conference on Computing Engineering and Design (ICCED) pp. 1 - 4
Main Authors Erfina, Adhitia, Tamanin, Aidah Sri, Listianawati, Sembiring, Falentino, Saepudin, Sudin, Melawati, Lesmana, CSA Teddy
Format Conference Proceeding
LanguageEnglish
Published IEEE 15.10.2020
Subjects
Fuzzy Mamdani
Machine Learning
Natural Language Processing
Sarcasm Detection
Sentiment analysis
Online AccessGet full text

Cover

More Information
Summary:Product reviews on e-commerce become very important for prospective buyers to determine the items to be bought. Unfortunately, the results of the product reviews are not accurate because of the meaningful reviews of sarcasm. Sarcasm is a form of intelligent speech where the speaker or writer says or writes the opposite of the meaning. Reviewers sometimes give 5 stars review but are accompanied by negative comments with the intention of being sarcastic. The number of reviews continues to grow so it is very difficult to make limits to classify whether the number of product reviews is positive or negative. Thus, the use of the Fuzzy Mamdani method is very important to determine the value of the majority of sentiments. This study presents a new approach to the field of sarcasm detection by combining the methods of sentiment analysis and Fuzzy Mamdani by comparing the allegedly sarcastic reviews of reviews from other buyers of similar products. The results of this study indicate that the greater the difference between positive and negative reviews, the greater the review is negative.
DOI:10.1109/ICCED51276.2020.9415787