Perbandingan metode vader, naive bayes, dan support vector machine untuk analisis sentimen ulasan produk somethinc pada website female daily

Agustin, Priscilla (2022) Perbandingan metode vader, naive bayes, dan support vector machine untuk analisis sentimen ulasan produk somethinc pada website female daily. Bachelor thesis, Universitas Pelita Harapan.

Preview	Text (Title) Title.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (63kB) | Preview
Preview	Text (Abstract) Abstract.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (309kB) | Preview
Preview	Text (ToC) ToC.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (853kB) | Preview
	Text (Chapter1) Chapter1.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (956kB)
	Text (Chapter2) Chapter2.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)
	Text (Chapter3) Chapter3.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (1MB)
	Text (Chapter4) Chapter4.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (2MB)
	Text (Chapter5) Chapter5.pdf Restricted to Registered users only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (188kB)
Preview	Text (Bibliography) Bibliography.pdf Available under License Creative Commons Attribution Non-commercial Share Alike. Download (403kB) | Preview
	Text (Appendices) Appendices.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial Share Alike. Download (2MB)

Abstract

Berbagai produk perawatan kulit dan wajah (skincare) tersedia untuk memenuhi kebutuhan orang-orang dalam menunjang penampilan fisik mereka, salah satunya adalah produk Somethinc. Somethinc adalah salah satu merek skincare lokal terbaik. Dalam memilih produk, konsumen seringkali mengamati ulasan yang diunggah oleh konsumen lainnya. Ulasan dari konsumen lain memiliki pengaruh yang besar terhadap minat beli seseorang. Salah satu wadah untuk menampung ulasan konsumen mengenai produk skincare adalah website Female Daily. Berdasarkan website Female Daily, dilakukan analisis sentimen terhadap produk Somethinc Niacinamide + Moisture Beet Serum untuk mengklasifikasikan sentimen menjadi positif, negatif, dan netral. Analisis sentimen dilakukan menggunakan metode VADER yang merupakan pendekatan berbasis kamus dan aturan, serta Naïve Bayes (NB) dan Support Vector Machine (SVM) yang merupakan pendekatan berbasis machine learning. Tahapan pada penelitian ini terdiri dari pengumpulan data, pra-pemrosesan data, pembobotan kata dengan metode TF-IDF, analisis sentimen menggunakan VADER, NB, dan SVM, dan evaluasi model. Penelitian ini dilakukan untuk membandingkan performa dari ketiga metode tersebut, serta mengetahui persepsi konsumen terhadap produk Somethinc. Hasil dari analisis sentimen menunjukkan bahwa algoritma SVM merupakan metode yang menghasilkan performa terbaik dengan nilai akurasi sebesar 86,03%, diikuti algoritma NB dengan nilai akurasi sebesar 85,04%, dan yang terakhir adalah VADER dengan nilai akurasi sebesar 48,78%. Hasil analisis sentimen dengan metode VADER menunjukkan bahwa mayoritas ulasan bersifat positif dan netral. Sementara itu, analisis sentimen dengan algoritma NB dan SVM menunjukkan bahwa ulasan secara keseluruhan bersifat positif. Hasil penelitian menunjukkan bahwa analisis sentimen menggunakan pendekatan berbasis machine learning lebih cocok digunakan./Various skincare products are available to fulfill people’s desire in improving their physical appearance, one of them is Somethinc. Somethinc is one of the best local skincare brands. When choosing a product, consumers often observe reviews uploaded by other consumers. Reviews from the other have a big influence on one’s buying interest. One of the platforms that accommodate consumer reviews about skincare products is Female Daily website. Based on the Female Daily website, the author conducted sentiment analysis towards Somethinc Niacinamide + Moisture Beet Serum product to classify sentiments into positive, negative, and neutral. Sentiment analysis was performed using VADER which is a lexicon and rule-based approach, as well as Naïve Bayes (NB) and Support Vector Machine (SVM) which are machine learning approaches. The steps in this research are data collecting, data pre-processing, term weighting using TF-IDF method, sentiment analysis using VADER, NB, and SVM, and model evaluation. The purpose of this research is to compare the performance of the three methods used, as well as to understand consumer perceptions about Somethinc. The result of the sentiment analysis shows that SVM algorithm produces the best performance with an accuracy of 86,03%, followed by NB algorithm with an accuracy of 85,04%, and lastly VADER with an accuracy of 48,78%. The result of sentiment analysis using VADER shows that most reviews are positive and neutral. Meanwhile, sentiment analysis using NB and SVM algorithms shows that most reviews are positive. The results of this research show that sentiment analysis using machine learning approach is more suitable.

Actions (login required)

View Item