Sarcasm Detection in Headline News using Machine and Deep Learning Algorithms

Download Edit this record How to cite View on PhilPapers
Abstract
Abstract: Sarcasm is commonly used in news and detecting sarcasm in headline news is challenging for humans and thus for computers. The media regularly seem to engage sarcasm in their news headline to get the attention of people. However, people find it tough to detect the sarcasm in the headline news, hence receiving a mistaken idea about that specific news and additionally spreading it to their friends, colleagues, etc. Consequently, an intelligent system that is able to distinguish between can sarcasm none sarcasm automatically is very important. The aim of the study is to build a sarcasm model that detect headline news using machine and deep learning and attempt to understand how a computer learns the patterns of sarcasm. The dataset used in this study was collected from Kaggle depository. We examined 21 algorithms of machine learning and one deep learning algorithm for detecting sarcasm in headline news. The evaluation metric used in this study are Accuracy, F1-measure, Recall, Precision, and Time needed for training and evaluation. The deep learning model achieved accuracy (95.27%), recall (96.62%), precision (94.15%), F1-score (95.37%) and time needed to train the mode (400 seconds), with loss of around 0.3398. However, the algorithm of machine learning that achieved the highest F1-Score is Passive Aggressive Classifier. It was the top classier for sarcasm detection among the machine learning algorithms with accuracy (95.50%), recall (96.09 %), precision (94.30%), F1-score (95.19%) and time needed to train the mode (0.31 seconds).
Categories
(categorize this paper)
PhilPapers/Archive ID
BARSDI-7
Upload history
Archival date: 2022-04-29
View other versions
Added to PP index
2022-04-29

Total views
98 ( #52,770 of 71,281 )

Recent downloads (6 months)
98 ( #7,192 of 71,281 )

How can I increase my downloads?

Downloads since first upload
This graph includes both downloads from PhilArchive and clicks on external links on PhilPapers.