Sistem Analisis Sentimen pada Ulasan Produk Menggunakan Metode Naive Bayes

Gunawan, Billy; Pratiwi, Helen Sasty; Pratama, Enda Esyudha

doi:10.26418/jp.v4i2.27526

Cited by 82 publications

(91 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Cuitan yang tersedia dari twitter dapat dimanfaatkan untuk mendapatkan infromasi (text mining) [1]. Sentimen akan diproses untuk menghasilkan informasi atau pola yang dapat diubah sebagai ilmu pengetahuan baru [2]. Tujuan dari pengklasifikasian sentimen adalah untuk menemukan kecendurangan pola opini dari ulasan pengguna untuk menganalisa dan menemukan masalah [3].…”

Section: Pendahuluanunclassified

Pebandingan Performa Naïve Bayes dan KNN pada Klasifikasi Teks Sentimen Jasa Ekspedisi

Putra¹,

Nugroho

2021

JOINTECS

View full text Add to dashboard Cite

This study aims to determine the performance of the naive bayes model and KNN (K-Nearest Neighbor) in classifying sentiment text expedition services. Twitter user reviews on @jntexpressid, @JNE_ID, and @posindonesia will be mined and classified based on positive, neutral, and negative sentiments Data was mined for two months from October1 to December 1 2020, the results obtained on the JNT data are 46.220, JNE 5.364, and Pos Indonesia 11.194. The raw data obtained will be cleaned and labeled before entering using TF-IDF for term weighting. The clean data obtained from the pre-processing of the text will be divided into 70% training data and 30% test data. The over-sampling method is used to balance the dataset and avoid overfitting training data. Naive ayes model and KNN there was a significant increase after the over-sampling method was implemented. The greatest accuracy in naive bayes is found in JNT data of 82%, while KNN on POS data is 71%. In KNN a high K value does not determine the accuracy results, it is proven that K = 2 has the highest accuracy compared to K = 4, K = 6, K = 8, K = 10 after the resampling technique is carried out.

show abstract

Section: Pendahuluanunclassified

Pebandingan Performa Naïve Bayes dan KNN pada Klasifikasi Teks Sentimen Jasa Ekspedisi

Putra¹,

Nugroho

2021

JOINTECS

View full text Add to dashboard Cite

show abstract

“…Klasifikasi Naïve Bayes merupakan klasifikasi yang bersifat supervised learning karena memiliki supervisor (manusia melakukan klasifikasi secara manual pada data yang digunakan dalam pelatihan) selaku pengajar dalam proses belajar atau learning. Selain itu, performansi Naïve Bayes memiliki waktu klasifikasi yang singkat sehingga mempercepat proses sistem analisis sentimen [12].…”

Section: Pemodelan Dataunclassified

Analisis Sentimen Zoom Cloud Meetings di Play Store Menggunakan Naïve Bayes dan Support Vector Machine

Herlinawati

Yuliani

Faizah

et al. 2020

Com, Engine, Sys, Sci

View full text Add to dashboard Cite

Aplikasi zoom cloud meetings yang mulai booming digunakan sekarang ini karena adanya pandemi virus corona, sehingga membuat semua kegiatan dilakukan secara virtual. Zoom cloud meetings merupakan aplikasi yang memiliki berbagai fitur termasuk video & audio conference. Pada penelitian ini penulis menggunakan metode Naïve Bayes dan Support Vector Machine dalam menganalisa label sentimen positif atau negatif pada ulasan para pengguna aplikasi zoom di Google Play Store. Jumlah dataset setelah prepocessing menjadi 1.007 record. Data hampir seimbang dengan label positif sebanyak 546 dan label negatif 461 ulasan. Evaluasi model menggunakan 10 fold cross validation diperoleh nilai akurasi dan nilai AUC dari masing-masing algoritma yaitu untuk NB nilai akurasi = 74,37% dan nilai AUC = 0,659. Sedangkan untuk algoritma SVM nilai akurasi = 81,22% dan nilai AUC = 0,886. Dalam penelitian ini dapat diketahui bahwa tingkat akurasi yang didapatkan algoritma Support Vector Machine (SVM) lebih unggul 6,85% dibandingkan algoritma Naïve Bayes (NB). Kata Kunci— Zoom Cloud Meetings, Google Play Store, Virus Corona, Naïve Bayes, Support Vector Machine. Abstract— Zoom cloud meetings application that began to boom is used today because of the corona virus pandemic, so that all activities are carried out virtually. Zoom cloud meetings is an application that has various features including video & audio conferencing. In this study the authors used the Naïve Bayes method and Support Vector Machine in analyzing positive or negative sentiment labels on the zoom users' reviews on the Google Play Store. The number of datasets after prepocessing is 1,007 records. The data is almost balanced with 546 positive labels and 461 negative labels. Evaluation of the model using 10 fold cross validation obtained accuracy values and AUC values from each algorithm, namely for NB, the accuracy value = 74.37% and the AUC value = 0.659. As for the SVM algorithm the accuracy value = 81.22% and the AUC value = 0.886. In this study it can be seen that the accuracy obtained by the Support Vector Machine (SVM) algorithm is 6.85% superior to the Naïve Bayes (NB) algorithm.

show abstract

“…This study classifies tweets from Twitter social media using sentiment analysis as a part of the text mining technique (Branz & Brockmann, 2018). There is much previous research of sentiment analysis from social media to get important information, among others: (1) sentiment analysis to analyze the satisfaction of citizens in public service (Akinboade, Kinfack, & Mokwena, 2012); (2) sentiment analysis of film review (Amrullah, Sofyan Anas, & Hidayat, 2020;Maylawati, Mudyawati, Wahisyam, & Maulana, 2021); (3) sentiment analysis to review product quality (Gunawan, Pratiwi, & Pratama, 2018); (4) sentiment analysis of government policy such as traffic control policy (Ruhyana, 2019) and bureaucratic service (Muliawaty, Alamsyah, Salamah, & Maylawati, 2019); and so on Based on that previous research, this study used the sentiment analysis technique to understand the disaster situation based on the response from the Government and public. The contribution of this research is to offer a method that can simultaneously identify responses from Government and society, which can help better understand disaster management.…”

Section: Introductionmentioning

confidence: 99%

Content Analysis of Social Media: Public and Government Response to COVID-19 Pandemic in Indonesia

Akbar

Kurniadi

Nurliawati³

2021

JSP

View full text Add to dashboard Cite

Nowadays, the use of social media to analyze disaster responses has become important. However, its application to support decision-making by the Government during disasters still present significant challenges. This article offers a complete analysis of the response of the public and the Government in dealing with the COVID-19 Pandemics in Indonesia. The content analysis uses to analyze the tweet post on Twitter to determine the public and government response. Data was collected from public and government tweets on Twitter and producing 11,578 community tweets from the public and 958 tweets from the government account. This data was collected from 2nd March until 15th April 2020. Public comments are sorted into six categories of comments, that is fate, logic, government mention, worry, scientist, and impression, while sentiments are classified as positive, negative, and neutral. Government comments are sorted into eight categories, namely information, education, operating, warnings, resources provision, volunteer recruitment, and rumors management. The results showed that the public encourages and supports the Government to cope with a pandemic think rationally and logically in dealing with this Pandemic. In addition, the study indicates that the Government has not used social media as a medium for communicating with the public. The quality of government response is not good, especially in the categories of information on operations, warnings, resources provision, recruitment of volunteers, and rumors management. The implication of this study suggests how the data might be useful for the Government in delivering information during the Pandemic.

show abstract

Sistem Analisis Sentimen pada Ulasan Produk Menggunakan Metode Naive Bayes

Cited by 82 publications

References 3 publications

Pebandingan Performa Naïve Bayes dan KNN pada Klasifikasi Teks Sentimen Jasa Ekspedisi

Pebandingan Performa Naïve Bayes dan KNN pada Klasifikasi Teks Sentimen Jasa Ekspedisi

Analisis Sentimen Zoom Cloud Meetings di Play Store Menggunakan Naïve Bayes dan Support Vector Machine

Content Analysis of Social Media: Public and Government Response to COVID-19 Pandemic in Indonesia

Contact Info

Product

Resources

About