Institutional Repository | Satya Wacana Christian University: Sentiment Analysis Model Based on Youtube Comment Using Support Vector Machine
Sentimen Analisis Model Berdasarkan Komentar Youtube
Menggunakan SUPPORT VECTOR MACHINE (SVM)
Fiktor Imanuel Tanesab1 Irwan Sembiring2, Hindriyanto Dwi Purnomo3
1,2,3
Fakultas Teknologi Informasi, Universitas Kristen Satya Wacana, Salatiga – Jawa Tengah, Indonesia.
1
[email protected], [email protected], [email protected]
ABSTRACT
Opinion mining or comment toward attitude evaluation, individual entity, are usually called
sentiment. Everyone is free to give opinion related with the present opinions on youtube. Hence
people have a free will to express their opinion regarding the performance. Due to the raise of
many critics that appear in a short amount of time, there a needs to conduct research on opinion
mining. In this research, opinion mining is applied on the peformance of Ahok as a governor.
The sentiment analysis is used to find a pattern or a certain character of Ahok. Support Vector
Machine is used to classified the opinion into positive class, neutral class and negative class.
1000 recorded data is used as a sample data. Preprocessing phase is needed before classifying the
data. The preprocessing phase consist of preprocessing the data, tokenizing, cleansing and
filtering. In order to determine the percentage of the class sentiment, Lexicon Based method is
used. The experiment shows that the proposed method are calculating the percentage weight in
this research had used Lexicon Based and Confusion Matrix to know the result of weighting
percentage of analysis to SVM. It had been found the result as follows : accuracy 84%, precision
91%, recall 80%, TP rate 91.1 and TN rate 44.8%.
Keywords: Youtube, Analysis Sentiment, Support Vector Machine, Opinion Mining, Lexicon
Based.
ABSTRAK
Opinion mining atau komentar terhadap penilaian sikap, entitas individu, bergerak begitu bebas,
hal seperti ini yang sering disebut sentimen. Di youtube semua orang bebas memberikan
pendapat atau beropinin terkait dengan opini – opini yang ada. Dengan adanya opinion yang
mengalir begitu cepat maka, perlu adanya penelitian terkait opininon mining untuk mengetahui
sejauh mana kinerja Gubernur Ahok. Analisa sentiment merupakan suatu cara untuk mengetahui
pola atau karakter dari Ahok. Pada penelitian ini digunakan metode Support Vector Machine
untuk mengetahui kinerja gubernur berdasarkan class positive, neutral dan negatitive, data yang
digunakan pada penelitian ini adalah 1000 record data. Untuk melakukan riset atas opini
masyarakat yang mengandung sentiment positive, neutral, atau negative, maka terdapat beberapa
preprocessing data yakni, tokenisasi, cleansing dan filtering, dan untuk menentukan persentase
class sentimen dengan metode Lexicon Based. Dari penelitian ini didapatkan nilai akurasi
sebagia berikut, accuracy 84%, precision 91%, recall 80%, TP rate 91.1 dan TN rate 44.8%.
Keywords : Youtube; Analysis sentiment; Support Vector Machine; Opinion mining; Lexicon
based.
Menggunakan SUPPORT VECTOR MACHINE (SVM)
Fiktor Imanuel Tanesab1 Irwan Sembiring2, Hindriyanto Dwi Purnomo3
1,2,3
Fakultas Teknologi Informasi, Universitas Kristen Satya Wacana, Salatiga – Jawa Tengah, Indonesia.
1
[email protected], [email protected], [email protected]
ABSTRACT
Opinion mining or comment toward attitude evaluation, individual entity, are usually called
sentiment. Everyone is free to give opinion related with the present opinions on youtube. Hence
people have a free will to express their opinion regarding the performance. Due to the raise of
many critics that appear in a short amount of time, there a needs to conduct research on opinion
mining. In this research, opinion mining is applied on the peformance of Ahok as a governor.
The sentiment analysis is used to find a pattern or a certain character of Ahok. Support Vector
Machine is used to classified the opinion into positive class, neutral class and negative class.
1000 recorded data is used as a sample data. Preprocessing phase is needed before classifying the
data. The preprocessing phase consist of preprocessing the data, tokenizing, cleansing and
filtering. In order to determine the percentage of the class sentiment, Lexicon Based method is
used. The experiment shows that the proposed method are calculating the percentage weight in
this research had used Lexicon Based and Confusion Matrix to know the result of weighting
percentage of analysis to SVM. It had been found the result as follows : accuracy 84%, precision
91%, recall 80%, TP rate 91.1 and TN rate 44.8%.
Keywords: Youtube, Analysis Sentiment, Support Vector Machine, Opinion Mining, Lexicon
Based.
ABSTRAK
Opinion mining atau komentar terhadap penilaian sikap, entitas individu, bergerak begitu bebas,
hal seperti ini yang sering disebut sentimen. Di youtube semua orang bebas memberikan
pendapat atau beropinin terkait dengan opini – opini yang ada. Dengan adanya opinion yang
mengalir begitu cepat maka, perlu adanya penelitian terkait opininon mining untuk mengetahui
sejauh mana kinerja Gubernur Ahok. Analisa sentiment merupakan suatu cara untuk mengetahui
pola atau karakter dari Ahok. Pada penelitian ini digunakan metode Support Vector Machine
untuk mengetahui kinerja gubernur berdasarkan class positive, neutral dan negatitive, data yang
digunakan pada penelitian ini adalah 1000 record data. Untuk melakukan riset atas opini
masyarakat yang mengandung sentiment positive, neutral, atau negative, maka terdapat beberapa
preprocessing data yakni, tokenisasi, cleansing dan filtering, dan untuk menentukan persentase
class sentimen dengan metode Lexicon Based. Dari penelitian ini didapatkan nilai akurasi
sebagia berikut, accuracy 84%, precision 91%, recall 80%, TP rate 91.1 dan TN rate 44.8%.
Keywords : Youtube; Analysis sentiment; Support Vector Machine; Opinion mining; Lexicon
based.