ANALISIS SENTIMEN TERHADAP LAYANAN PROVIDER TELKOMSEL PADA JEJARING SOSIAL TWITTER MENGGUNAKAN ALGORITMA SUPPORT VECTOR MACHINE
Total View This Week0
Total View This Week0
Institusion
Universitas Pembangunan Nasional Veteran Jakarta
Author
Roihan Mufli Arjuna, .
Subject
QA75 Electronic computers. Computer science
Datestamp
2021-12-21 07:54:08
Abstract :
Twitter is one of social media is used to have complains, critics, and suggestions from customer by established companies. One of them is Telecommunication Company in Indonesia, it is PT Telekomunikasi Selular or so-called Telkomsel. This research will study sentiment analysis customer on Twitter by tagging Telkomsel account with username @Telkomsel. Fetched data is data tweet, they were fetched by using Twitter API (Application Programming Interface).Then data tweet was labeled positive and negative by annotator. After that, preprocessing data such as cleaning, case folding, normalization, stopword removal, stemming, and tokenization then term weighting with Term Frequency – Inverse Document Frequency (TF-IDF). Before classification process, data is oversampled by using SMOTE technique because there is 614 positive and 2214 negative which mean it is imbalanced data. After oversampling, data become balanced it is 2214 positive and 2214 negative. Then data was divided to 70% as data training and 30% as data testing randomly. Algorithm is used to classify is support vector machine. Classification result is accuracy has 0.93, precision has 0.91, recall has 0.96, specificity has 0.91, F1 score has 0.93.
Twitter is one of social media is used to have complains, critics, and suggestions from customer by established companies. One of them is Telecommunication Company in Indonesia, it is PT Telekomunikasi Selular or so-called Telkomsel. This research will study sentiment analysis customer on Twitter by tagging Telkomsel account with username @Telkomsel. Fetched data is data tweet, they were fetched by using Twitter API (Application Programming Interface).Then data tweet was labeled positive and negative by annotator. After that, preprocessing data such as cleaning, case folding, normalization, stopword removal, stemming, and tokenization then term weighting with Term Frequency – Inverse Document Frequency (TF-IDF). Before classification process, data is oversampled by using SMOTE technique because there is 614 positive and 2214 negative which mean it is imbalanced data. After oversampling, data become balanced it is 2214 positive and 2214 negative. Then data was divided to 70% as data training and 30% as data testing randomly. Algorithm is used to classify is support vector machine. Classification result is accuracy has 0.93, precision has 0.91, recall has 0.96, specificity has 0.91, F1 score has 0.93.
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
