Analisis Data Mining Klasifikasi Berita Hoax COVID 19 Menggunakan Algoritma Naive Bayes

Abstract

The rapid dissemination of information along with the rapid development of technology along with the massive speed of electronic media and the internet. But the rapid spread of news cannot guarantee that the information and news that we get can be validated from valid sources. Based on data released by Kominfo at the end of 2021, there were 1773 hoax news that were successfully clarified from the hoax news. Then during the Covid-19 pandemic itself, there were various hoaxes circulating in the community. Throughout 2021, the Ministry of Communications and Informatics discovered as many as 723 hoaxes about Covid-19. Based on the background above, the researchers and previous studies have discussed hoax detection in various fields. Such as, fraud detection in online writing style [1], classification of hoax news based on machine learning [3] and the application of nave Bayes and PSO algorithms for classification of hoax news on social media [4]. From here the researchers tried to carry out experiments on the nave Bayes classification algorithm to classify hoax covid 19 news. Based on the results of research that has been done, the nave Bayes model and cross validation can classify hoax news well, the resulting accuracy is 86.3% where 80-90% included in the good classification criteria. The data that is predicted to be incorrect is also not too much from a total of 300 datasets, only 41 are declared incorrect in labeling less than 2% of the total dataset, so it can be concluded that this model can be used as a reference if you want to proceed to a more complex prediction model, for example the model prediction using web-based machine learning.

Penggunaan Algoritma Naïve Bayes Dalam Menganalisis Sentimen Ulasan Aplikasi Adakami Di Google Play Store

“…Teks yang telah diproses disebut sebagai istilah, dan setiap istilah dalam dokumen dapat memiliki berbagai makna, termasuk beberapa yang tidak relevan. Menghilangkan istilah yang tidak relevan dari dokumen merupakan bagian dari proses tersebut, sementara istilah yang bermakna tetap ada dan berfungsi sebagai representasi dokumen [8]. Menghapus kata bertujuan untuk menghilangkan istilah-istilah yang tidak perlu, sehingga mengurangi dimensi dataset [9].…”

Sentiment Analysis of Google Play Store User Reviews on Digital Population Identity App Using K-Nearest Neighbors

“…Text Mining is the search for unknown information using automated data extraction from large amounts of unstructured text [18]. The purpose of text mining is to analyze the opinions, feelings, judgments, attitudes, assessments, feelings of a person to find out whether a topic, service, organization or person is related to a particular activity [19].…”

Hoax Detection News Using Naïve Bayes and Support Vector Machine Algorithm

“…Previous studies have discussed the detection of hoaxes in various information. Like Covid 19 [3] conducted experiments on the Naïve Bayes classification algorithm to classify hoax news Covid 19. Based on the results of research that has been carried out, the naïve Bayes model and cross validation can classify hoax news well, the resulting accuracy is 86.3% where 80-90% fall into the criteria of good classification.…”