文章基本信息

标题：Sarcasm Detection in Amazon Product Reviews
本地全文：下载
作者：Sahil Jain ; Ashish Ranjan ; Dipali Baviskar 等
期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2018
卷号：9
期号：3
页码：108-111
出版社：TechScience Publications
摘要：Sarcasm detection: Sarcasm is defined as the use ofirony to mock or convey contempt. It is a feature of naturallanguage where the literal meaning of one’s language is quiteopposite of the implied meaning. On e-commerce websitessuch as amazon, many times customers make the use ofsarcasm in their reviews in an attempt to criticize the product.With the help of sarcasm detection, products can be classifiedinto the relevant categories with more accuracy. Methodology:A corpus of ironic and regular reviews is used for the purposeof this experiment. The data is extracted into pythondictionaries using the module such as beautiful soup.Sentiment features, lexical features and parts of speechfeatures are extracted from the training and testing data usingNLTK and TextBlob. Various classifiers are trained with thefeatures of the training set. The results are calculated based onthe predictions of the testing set. Results: Reviews from thetesting set are classified as sarcastic/ironic or regular. Basedon the observations, the accuracy, precision, recall and f-scoreare calculated. Hence, the performance of our experimentalsetup is evaluated. Applications: Opinion mining,recommendations and advertisement systems.
关键词：Irony; Natural Language Processing; Opinion;Mining; Sarcasm; Sarcasm Detection; Sentiment Analysis; SVM;Introduction