Implementasi Metode K-Nearest Neighbor Untuk Klasifikasi Tweet Hate Speech Pada Media Sosial Twitter
Total View This Week0
Total View This Week0
Institusion
Politeknik Negeri Jember
Author
Wulandari, Franciska
Subject
458 - Teknik Informatika
Datestamp
2022-03-31 02:17:20
Abstract :
Twitter is a microblogging - based social media. Microblogging is a type of social media that facilitates users to write and publish activities or opinions freely. Twitter not only has a positive impact, but also has a negative impact. The impact is the emergence of various types of crime violations, for example hate speech. Of course, it takes the help of a linguist to identify hate speech where it can take a long time to be identified using the system. This study uses the K-Nearest Neighbor method, 1000 total of data divided into 700 as training data and 300 as test data. with the classification of religious hate speech, racial and neutral hate speech. The results of the testing process using the confusion matrix obtained the accuracy of the training data by 70%, 80%, and 90%, namely 98.5%, 98.3%, and 98.5% with a value of k=7. The classification of the majority of testing data is 33% religious hate speech, 17% racial hate speech and 50% classified as neutral.
Twitter is a microblogging - based social media. Microblogging is a type of social media that facilitates users to write and publish activities or opinions freely. Twitter not only has a positive impact, but also has a negative impact. The impact is the emergence of various types of crime violations, for example hate speech. Of course, it takes the help of a linguist to identify hate speech where it can take a long time to be identified using the system. This study uses the K-Nearest Neighbor method, 1000 total of data divided into 700 as training data and 300 as test data. with the classification of religious hate speech, racial and neutral hate speech. The results of the testing process using the confusion matrix obtained the accuracy of the training data by 70%, 80%, and 90%, namely 98.5%, 98.3%, and 98.5% with a value of k=7. The classification of the majority of testing data is 33% religious hate speech, 17% racial hate speech and 50% classified as neutral.
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
