(MIB - S3) - Transiska, Febriawan, Hasan [2024-04-30]

Hasan, Firman Noor (2024) (MIB - S3) - Transiska, Febriawan, Hasan [2024-04-30]. MIB: Jurnal Media Informatika Budidarma, 8 (2). pp. 1077-1086. ISSN 2548-8368

	Text 1. Cover.pdf Download (83kB)
	Text 2. Editorial Team.pdf Download (277kB)
	Text 3. Daftar Isi.pdf Download (544kB)
	Text 4. Artikel.pdf Download (1MB)
	Text 5. Bukti Hasil Turnitin.pdf Download (2MB)
	Text 6. LOA - Transiska, Febriawan, Hasan.pdf Download (168kB)

Abstract

Chatbots can assist by fostering collaboration between users and companies or organizations. In today's world, artificial intelligence through chatbots has become one of the mainstays in handling various problems. One of the well-known types of chatbots is ChatGPT, which utilizes NLP (Natural Language Processing) technology to understand and respond to user queries and requests. ChatGPT, as a device that is very easy to use for all circles, has a fairly simple interface but does not cause boredom, and the speed in responding to commands given, this is an added value of ChatGPT. Despite the myriad of conveniences offered, ChatGPT also raises concerns on the negative side. The negative side is that there are many concerns that arise, starting from the rampant spread of hoaxes and misunderstandings on social media. The advantages and disadvantages that have been explained above, researchers are encouraged to find out the truth from the public's response regarding ChatGPT more deeply so that this sentiment analysis research is made. Moreover, research related to sentiment analysis can be said to be quite an answer to the confusion of public responses outside related to ChatGPT. This research also starts from the process of Data retrieval on Twitter social media using Rapidminer, in this process the researcher uses the Twitter API token on the Rapidminer application so that it can be obtained. The data that has been obtained is then cleaned through the preprocessing process using the features available in Rapidminer, the result of this process is that the data becomes clean. After being cleaned through preprocessing, it is then labeled as positive or negative which will later be classified by the Naïve Bayes algorithm. This classification aims to divide between positive sentiment and negative sentiment. After performing classification, the data is then evaluated using a confusion matrix and the results are obtained with an accuracy value of 96.55%, a precision value of 89.19%, and a recall value of 95.18%.

Item Type:	Article
Subjects:	T Technology > T Technology (General)
Divisions:	Fakultas Teknik > Teknik Informatika
Depositing User:	Mr Firman Noor Hasan
Date Deposited:	03 Jun 2024 02:17
Last Modified:	03 Jun 2024 02:17
URI:	http://repository.uhamka.ac.id/id/eprint/34340

Actions (login required)

View Item