JCT at SemEval-2022 Task 6-A: Sarcasm Detection in Tweets Written in English and Arabic using Preprocessing Methods and Word N-grams

Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022), pages 1031 - 1038 July 14-15, 2022 ©2022 Association for Computational Linguistics JCT at SemEval-2022 Task 6-A: Sarcasm Detection in Tweets Written in English and Arabic using Preprocessing Methods and Word N-grams Yaakov HaCohen-Kerner, Matan Fchima, Ilan Meyrowitsch Department of Computer Science, Jerusalem College of Technology, Lev Academic Center 21 Havaad Haleumi St., P.O.B. 16031, 9116001 Jerusalem, Israel kerner@jct.ac.il, matanf1992@gmail.com, meyrowitsch@gmail.com Abstract In this paper, we describe our submissions to SemEval-2022 contest. We tackled subtask 6-A - “iSarcasmEval: Intended Sarcasm Detection In English and Arabic – Binary Classification". We developed different models for two languages: English and Arabic. We applied 4 supervised machine learning methods, 6 preprocessing methods for English and 3 for Arabic, and 3 oversampling methods. Our best submitted model for the English test dataset was an SVC model that balanced the dataset using SMOTE and removed stop words. For the Arabic test dataset, our best submitted model was an SVC model that preprocessed removed longation. 1 Introduction The rapid development of various types of social networks allows the development of increasingly offensive language in general and sarcasm in particular. Sarcasm is the use of words that mean the opposite of what you want to say especially to insult someone, show irritation, or be funny1. Sarcastic language can harm individuals or groups of people and may cause harmful effects on society. Thus, it is important to develop high- quality systems capable of detecting sarcastic expressions automatically. Sarcasm detection is a difficult task not only for a computer but even for a human being. High- quality performance of this task requires understanding the context of the situation, the relevant culture, and in some cases the specific 1 www.merriam-webster.com. Retrieved 9-Feb-2022. issue or people involved in this situation (Maynard and Greenwood, 2014). Moreover, the noisy nature of social media texts especially Twitter messages make the detection task even harder. Therefore, it is an interesting and challenging task to detect sarcasm using supervised machine learning (ML) methods and natural language processing (NLP) tools. Furthermore, Rosenthal et al. (2014) show a significant drop in sentiment polarity classification performance when processing sarcastic tweets, compared to non-sarcastic ones. In this paper, we describe our research and participation in subtask 6-A for sarcasm detection in tweets written in two languages: English and Arabic. The full description of task 6 in general and 6-A in particular including the datasets and the participating teams is given in Abu Farha et al. (2022). The structure of the rest of the paper is as follows. Section 2 introduces a background concerning sarcasm detection, text preprocessing, and text classification with imbalanced classes. Section 3 describes subtask 6-A and its datasets. In Section 4, we present the submitted models and their experimental results. Section 5 summarizes and suggests ideas for future research. 2 Related Research Most prior textual sarcasm detection datasets have been annotated by using either manual labeling or a weak supervision method. In the first approach, sarcasm labels are …

Content truncated to 500 words