Implementasi Text Mining Untuk Pengelompokan Ulasan Pelanggan E-Commerce Berdasarkan Topik Ulasan Menggunakan Algoritma Hierarchical Agglomerative Clustering

Manzil, Li'Izza Diana (2019) Implementasi Text Mining Untuk Pengelompokan Ulasan Pelanggan E-Commerce Berdasarkan Topik Ulasan Menggunakan Algoritma Hierarchical Agglomerative Clustering. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06111540000049-Undergraduate_Theses.pdf]
Preview
Text
06111540000049-Undergraduate_Theses.pdf

Download (409kB) | Preview
[thumbnail of 06111540000049-Undergraduate_Theses.pdf]
Preview
Text
06111540000049-Undergraduate_Theses.pdf

Download (409kB) | Preview
[thumbnail of 06111540000049-Undergraduate-Thesis.pdf]
Preview
Text
06111540000049-Undergraduate-Thesis.pdf

Download (1MB) | Preview

Abstract

Pesatnya perkembangan teknologi membuat masyarakat lebih mudah melakukan kegiatan apapun. Misalnya dalam kegiatan bisnis. Kini sebagian besar masyarakat lebih suka melakukan pembelian melalui internet, karena lebih mudah dalam transaksinya. Hal itu mengakibatkan banyak munculnya e-commerce di Indonesia. Namun, pembelian barang melalui e-commerce terkadang tidak sesuai dengan apa yang diinginkan, baik secara pelayanan maupun barang yang diperoleh. Oleh karena itu, banyak e-commerce yang memberikan fasilitas pemberian ulasan mengenai pelayanan ataupun barang. Ulasan yang diberikan pelanggan sangat beragam. Dengan demikian, perlu adanya suatu penelitian dengan tujuan mengelompokkan ulasan untuk memudahkan penjual dan calon pelanggan e-commerce dapat mengetahui kualitas barang melalui aspek yang diulas oleh pelanggan-pelanggan sebelumnya. Metode yang digunakan pada penelitian ini adalah clustering dengan algoritma Hierarchical Agglomerative Clustering (HAC). Untuk menentukan jumlah cluster pada ulasan dilakukan pemotongan jarak cluster pada titik threshold tertentu. Threshold yang digunakan pada proses clustering adalah 0.94, 0.95 dan 0.96 dengan masing-masing memiliki hasil 15, 8 dan 4. Dari ketiga proses clustering tersebut hasil terbaik dimiliki proses clustering dengan threshold 0,96 dengan nilai presisi sebesar 85.2 %.
=================================================================================================================================
The rapid development of technology has made it easier for people to carry out any activities. For example, in business activities. Now most people like to make purchases through the internet, because it is easier in the transaction. That requires many e-commerce transfers in Indonesia. However, the purchase of goods through e-commerce is issued not in accordance with what is desired, both services and goods obtained. Therefore, many e-commerce services provide facilities for reviewing goods services. Customer reviews are very diverse. Thus, there is a need for a study to classify and classify assessments for providers and prospective e-commerce customers to find out what aspects are reviewed by customers. The method that will be used in this study is cluster with the Hierarchical Agglomerative Cluster (HAC) algorithm. To determine the number of clusters in the review, the cluster distance is cut at a certain threshold point. The threshold used in the cluster process is 0.94, 0.95 and 0.96 with each having results of 15, 8 and 4. Based on the three clustering processes the best results have a clustering process with a threshold of 0.96 with a precision value of 85.2%.

Item Type: Thesis (Other)
Additional Information: RSMa 005.1 Man i-1 2019
Uncontrolled Keywords: E-Commerce, HAC, Ulasan
Subjects: Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
Q Science > QA Mathematics > QA278 Cluster Analysis. Multivariate analysis. Correspondence analysis (Statistics)
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Chemistry > 47201-(S1) Undergraduate Thesis
Depositing User: Li'Izza Diana Manzil
Date Deposited: 03 Oct 2024 08:03
Last Modified: 03 Oct 2024 08:03
URI: http://repository.its.ac.id/id/eprint/66198

Actions (login required)

View Item View Item