Sexism Speech Detection On The Instagram Platform With Naïve Bayes Algorithm

Narendragharini, Kamila and -, Endang Wahyu Pamungkas, S.Kom, M.Kom. (2024) Sexism Speech Detection On The Instagram Platform With Naïve Bayes Algorithm. Skripsi thesis, Universitas Muhammadiyah Surakarta.

	PDF (Naskah Publikasi) Naskah Publikasi.pdf Download (380kB)
	PDF (Surat Pernyataan Tidak Publikasi) Surat Pernyataan.pdf Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Sexism on social media, particularly on Instagram, has become a widespread problem in recent years. Despite Instagram's policies and guidelines against hate speech and discrimination, users continue to face gender-based discrimination, harassment, and objectification on the Instagram platform. This can manifest in a variety of aspects, including the sexualization of women's bodies, derogatory remarks about women, and the perpetuation of gender stereotypes. Women, in particular, are frequently become the targets of these sexist behaviors, which have a negative impact on their mental health and self-esteem. The study used Instagram because of the availability of public comments for individual opinions, with the goal of developing a program capable of detecting sexist language in Instagram comments section. The researchers used the Naive Bayes algorithm and multiple classification types, such as Gaussian, Bernoulli, and Multinomial, to achieve this study. This method accurately determines whether a word in a sentence is sexist or non-sexist. This classification process was carried out using the Python programming language. at classification stage, The data used originates from Instagram comments with total of 814, consisting of 177 comments labeled as sexism and 637 comments labeled as non-sexism. The dataset is then split into 80% for training and 20% for testing purposes, where 651 data as training and 163 as testing data. The results of the testing process with data without preprocessing have better accuracy and the multinomial method has the highest accuracy of 88% using data without preprocessing.

Item Type:	Thesis (Skripsi)
Uncontrolled Keywords:	sexism, women, Instagram, naïve bayes, social media
Subjects:	T Technology > TZ Technical Information
Divisions:	Fakultas Ilmu Komunikasi dan Informatika > Teknik Informatika
Depositing User:	KAMILA NARENDRAGHARINI
Date Deposited:	06 Feb 2024 08:34
Last Modified:	06 Feb 2024 08:34
URI:	http://eprints.ums.ac.id/id/eprint/121447

Actions (login required)

View Item