<p>Masyarakat modern dengan kesibukan sehari-harinya tentu akan mendapat tekanan emosional yang cukup tinggi. Hal yang dilakukan untuk meredakan emosi tersebut adalah salah satu dengan mendengarkan musik. MOODSIC merupakan sebuah aplikasi yang dapat memutar musik sesuai dengan ekspresi wajah pengguna. Aplikasi MOODSIC dibangun menggunakan mesin pengenalan ekspres wajah berbasis DCT dan LDA serta algoritma klasifikasi statistik. Berdasarkan hasil pengujian secara <em>off-line</em> mesin pengenalan ekspresi wajah berhasil memberikan performa yang baik, dengan akurasi sebesar 100% untuk data masukkan terdiri atas fitur DCT 144 elemen, 6 eigen vektor LDA dan klasifikasi statistik jenis LDA. Mesin pengenalan ekspresi wajah memerlukan waktu pengenalan yang pendek yaitu 1 milidetik. Secara <em>real-time</em> MOODSIC memberikan hasil yang cukup baik dengan akurasi pengenalan ekspresi sebesar 91.51% atau dengan tingkat kesalahan pengenalan 9.49%.</p><p> </p><p><em><strong>Abstract</strong></em></p><p><em><strong></strong></em><em>Modern society lifestyles face many activities every day, which make people receive a fairly high emotional stress. To reduce such kind of emotions can be treated by listening music. MOODSIC is an application that can play music according to the user's face expression. MOODSIC is developed using face expression recognition machine based on DCT, LDA and statistical classification algorithm. Based on offline testing result, face expression recognition machine successfully give good performance with accuracy of 100% when DCT features are 144 elements, 6 eigen vectors of LDA and kind of statistical classifier is LDA. The face expression recognition engine took shorter time to classification about 1 milliseconds. MOODSIC also give good performance with the accuracy of expression recognition about 91.51% or recognition error of 9,49% for real-time testing.</em></p>
There has been a significant increase in communication activities between internet users in online media due to the increase in social media users. For instance, Twitter users may send messages via their tweets. However, tweets can also contain negative meanings. Therefore, it deserves special attention as it has the potential to contain hate speech. Even the government deems it necessary to publish regulations to deal with hate speech cases such as the Information and Electronic Transactions Law (ITE Law) issued in 2018 Article 28 paragraph 2 of the Hate Speech. Machine Learning (ML) is one of the techniques that can be used in identifying patterns. There are various types of data that ML can be applied to, including text (known as Text Analytic). Previous research has used the Support Vector Machine (SVM) method to identify hate speech on Twitter text with more than one label (multilabel). The purpose of this study was to identify hate speech on Twitter with a label of more than one (multilabel) via Convolutional Neural Network (CNN). The study obtained the best CNN model with an accuracy of 98.76% from the multi-label dataset on hate speech in Indonesian texts
Social media is a platform that allows users to express themselves freely including spreading hate speech content. The government has issued the regulation in the UU ITE to handle and prevent hate speech on social media. The research was also conducted using the Bi-LSTM to classify the text into hate speech or not. Another research was purposed to detect hate speech and its categories using Bi-GRU. However, the performance of the model Bi-GRU is still lower than Bi-LSTM with an accuracy of 86.44% and 96.44%. Therefore, this study aims to build a model that can detect hate speech and its categories. The research offers Bi-LSTM as a classification model and IndoBERT as a tokenization model. The dataset used is a public dataset containing 13 thousand tweets. As a result, the best model obtained is using 20 epochs, 192 batch sizes, 1 layer Bi-LSTM with 40 nodes, and applying class weighing in the optimization process. The pre-train model from IndoBERT that is used to support the performance of the model in classifying is "indobenchmark/indobert-large-p2". The performance given by the purposed model is very good with an average accuracy, precision, and recall of 97.66%, 96.50%, and 85.25%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.