Rahmaningsih, Anggie (2023) Analisis Brand Sentiment Berdasarkan Review di Media Sosial pada Brand Smartphone Indonesia dengan Menggunakan Text Mining. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 09111940000050_Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
09111940000050_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (3MB) | Request a copy |
Abstract
Ketatnya persaingan brand smartphone di Indonesia mengharuskan perusahaan untuk mempertahankan kualitas produk mereka. Seberapa baik kualitas produk yang diterima memengaruhi kepuasan pelanggan. Salah satu bukti kepuasan pelanggan adalah sentiment pelanggan terhadap produk brand smartphone yang diutarakan di media sosial. Kolom komentar di Youtube menjadi tempat paling mudah untuk menyampaikan pendapat. Penelitian ini bertujuan untuk menganalisis sentiment dari 6 brand smartphone dengan market share tertinggi di Indonesia serta mengetahui atribut apa yang memengaruhi sentiment tersebut. Data penelitian diambil dari review berupa komentar positif maupun negatif di media sosial Youtube. Metode text mining yaitu sentiment analysis dan topic modeling digunakan untuk mengetahui bagaimana sentiment dari brand smartphone yang diteliti. Algoritma Support Vector Machine (SVM) digunakan dalam melakukan sentiment analysis penelitian ini. Latent Dirichlet Allocation (LDA) digunakan untuk mengetahui topik yang dibicarakan dalam topic modeling. Selain itu, word cloud digunakan untuk menggambarkan kata yang sering dibahas pada keseluruhan review. Penelitian ini menemukan bahwa enam brand smartphone yang diteliti memiliki sentiment positif lebih besar daripada sentiment negatif. Besar persentase sentiment tiap brand smartphone berbeda-beda. Topik yang dibicarakan oleh pengguna pada tiap brand berbeda. Sepuluh topik yang sering dibicarakan adalah pada brand smartphone di Indonesia adalah kamera, harga, layar, desain, dynamic island, game, upgrade, chipset, charger, dan speaker. Penelitian selanjutnya diharapkan dapat menutupi keterbatasan, kekurangan serta memperluas temuan dari penelitian ini dengan mengikuti tren smartphone pada saat itu. Teknik atau metode analisis dapat menggunakan algoritma lain yang memiliki akurasi lebih tinggi.
====================================================================================================================================
The intense competition for smartphone brands in Indonesia requires companies to maintain the standard of their goods. How good the quality of the product received affects customer satisfaction. One proof of customer satisfaction is customer sentiment towards smartphone brand products expressed on social media. The comment column on Youtube is the easiest place to express your opinion. This research seeks to examine the sentiments of 6 smartphone brands with the highest market share in Indonesia and discover what attributes influence these sentiments. The research data was taken from reviews in the form of positive and negative comments on Youtube social media. The text mining method, namely sentiment analysis and topic modeling, is used to find out how the sentiment of the smartphone brand under study is. The Support Vector Machine (SVM) technique is employed in this work to carry out sentiment analysis. In topic modeling, Latent Dirichlet Allocation (LDA) is utilized to determine the subjects that have been discussed. In addition, the word cloud is used to describe words that are often discussed in the entire review. This study found that the six smartphone brands studied had greater positive sentiment than negative sentiment. The percentage of sentiment for each smartphone brand is different. The topics discussed by users for each brand are different. Ten topics that are often discussed on smartphone brands in Indonesia are cameras, prices, screens, designs, dynamic islands, games, upgrades, chipsets, chargers, and speakers. Future research is expected to cover the limitations, deficiencies and expand the findings of this study by following the trend of smartphones at that time. Analysis techniques or methods can use other algorithms that have higher accuracy.
Actions (login required)
 |
View Item |