Fitriyah, Ayu Ni`matul (2019) Analisis Sentimen pada Teks Ulasan Pelanggan E-Commerce Berdasarkan Rating Menggunakan N-Gram dan Neuro-Fuzzy. Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
0611154000013-Undergraduated-Thesis.pdf Download (3MB) | Preview |
Abstract
Perkembangan e-commerce di Indonesia cukup pesat bahkan diprediksi akan tumbuh sebesar 144.5% pada tahun 2022. Dalam perkembangannya e-commerce memberikan fasilitas pada pelanggan untuk memberikan pendapat mereka mengenai produk yang mereka jual berupa ulasan dan nilainya. Ulasan memiliki pengaruh dalam penentuan keputusan pembelian di masa depan dan memiliki hubungan dengan kepercayaan pelanggan, namun terkadang terjadi kesalahan berupa ketidak cocokan antara nilai dan ulasan yang diberikan dan mengakibatkan nilai ulasan yang tersimpan tidak akurat sehingga diperlukan suatu sistem untuk melakukan prediksi nilai dari suatu teks ulasan agar suatu ulasan memiliki kecocokan dengan nilai yang diberikan. Pada penelitian ini dilakukan prediksi sentimen teks ulasan berdasarkan rating secara otomatis menggunakan FCM-ANFIS untuk klasifikasi, penambahan metode N-gram pada pra proses data, dan pengurangan dimensi dengan truncated SVD. Penggunaan n-gram meliputi unigram, bigram, dan kombinasi unigram bigram. Pada pengujian sistem dengan 4 kasus diperoleh akurasi untuk unigram 0.657, 0.83 0.919, 0.87 sedangkan bigram 0.64, 0.75, 0.79, 0.77 dan kombinasi 0.69, 0.81, 0.929, 0.889. Hasil pengujian ini menunjukkan bahwa penggunaan kombinasi unigram dan bigram mampu meningkatkan hasil akurasi dari prediksi sentimen teks ulasan.
==================================================================================================================================
E-commerce in Indonesia is growing rapidly and projected will increase up to 144,5% in 2020. In its development, e-commerce provides facilities to customer to give their opinions about the product sold by them in the form of review and rating. Reviews have an influence in determining future purchasing and have a relationship with customer trust, but sometimes an error occurs between the rating and review given by customers which affects the stored rating being inaccurate. A system is needed to predict the rating of a review text automatically so the review match with the rating. This research is prediction of textual review sentiment based on rating using FCM-ANFIS for classification model, addition of N-gram in pre-processing data, and using truncated SVD for dimensionality reduction. N-gram that used are unigram, bigram, and combination both of them. The testing scenario use 4 cases that have results unigram 0.657, 0.83 0.919, 0.87 for unigram 0.64, 0.75, 0.79, 0.77 for bigram, and 0.69, 0.81, 0.929, 0.889 for combination both of them. The result indicates that use combination of unigram and bigram is able to improve the accuracy of sentiment prediction of textual review based on rating.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | ulasan, nilai ulasan, analisis sentimen, N-Gram, Fuzzy C-Means, Adaptive Neuro Fuzzy Inference System |
| Subjects: | Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) Q Science > QA Mathematics > QA9.64 Fuzzy logic |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis |
| Depositing User: | Ayu Ni'matul Fitriyah |
| Date Deposited: | 26 Sep 2024 02:37 |
| Last Modified: | 26 Sep 2024 02:37 |
| URI: | http://repository.its.ac.id/id/eprint/66323 |
Actions (login required)
 |
View Item |