Tambunan, Norman Ernest Elmosaputra and Asri, Yessy and Kuswardani, Dwina (2025) PENERAPAN NAMED ENTITY RECOGNITION BERBASIS MACHINE LEARNING (CONDITIONAL RANDOM FIELDS) PADA ULASAN PENGADUAN DAN PELAYANAN PLN MOBILE. Diploma thesis, ITPLN.
202031061_Norman Ernest Elmosaputra Tambunan__NORMAN Ernest.pdf
Restricted to Registered users only
Download (4MB)
Abstract
Tujuan dari penelitian ini adalah untuk mengembangkan model analisis sentimen yang dapat mengidentifikasi dan mengklasifikasikan entitas penting dalam ulasan pengguna aplikasi PLN Mobile. Dengan menggunakan teknik web scraping, data ulasan sebanyak 24.000 diambil dari Google Play Store selama periode 2022-2024. Proses preprocessing dilakukan untuk membersihkan dan menyiapkan data, termasuk tokenisasi, lemmatisasi. Dalam ekstraksi fitur, dilakukannya (Part Of Speech)POS Tagging, (Begin, Inside, Last Outside)BILOU, serta Word Embedding.. Model Conditional Random Fields (CRF) diterapkan untuk melakukan Named Entity Recognition (NER) pada data yang telah diproses. Hasil evaluasi menunjukkan bahwa model CRF mampu mencapai akurasi tinggi dengan score 94% dalam mengidentifikasi entitas, meskipun terdapat beberapa kelas yang memiliki performa lebih rendah. Penelitian ini memberikan kontribusi signifikan dalam meningkatkan pemahaman dan pengelolaan umpan balik pengguna, serta menawarkan solusi untuk meningkatkan kualitas pelayanan PLN melalui analisis data yang lebih efisien.
The purpose of this research is to develop a sentiment analysis model that can identify and classify important entities in user reviews of the PLN Mobile application. Using web scraping technique, 24,000 review data were retrieved from Google Play Store during the period 2022-2024. Preprocessing is done to clean and prepare the data, including tokenization, lemmatization. In feature extraction, (Part Of Speech) POS Tagging, (Begin, Inside, Last Outside) BILOU, and Word Embedding are performed. Conditional Random Fields (CRF) model is applied to perform Named Entity Recognition (NER) on the processed data. The evaluation results show that the CRF model is able to achieve high accuracy with a score of 94% in identifying entities, although there are some classes that have lower performance. This research makes a significant contribution in improving the understanding and management of user feedback, as well as offering solutions to improve the quality of PLN services through more efficient data analysis.
| Item Type: | Thesis (Diploma) |
|---|---|
| Uncontrolled Keywords: | Named Entity Recognition, Corpus, Conditional Random Fields, Confusion Matrix Named Entity Recognition, Conditional Random Fields, Confusion Matrix |
| Subjects: | Skripsi Bidang Keilmuan > Teknik Informatika |
| Divisions: | Fakultas Telematika Energi > S1 Teknik Informatika |
| Depositing User: | Sudarman |
| Date Deposited: | 09 Oct 2025 06:10 |
| Last Modified: | 09 Oct 2025 06:10 |
| URI: | https://repository.itpln.ac.id/id/eprint/1990 |
