Open AccessThe accuracy of transfer learning using neural network method for sentiment analysis problem on Indonesian tweets
Author(s) -
A. N. Augustizhafira,
Hendri Murfi,
Gianinna Ardaneswari
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1725/1/012015
Subject(s) - computer science , trigram , artificial intelligence , transfer of learning , sentiment analysis , artificial neural network , machine learning , feature selection , bigram , feature (linguistics) , classifier (uml) , philosophy , linguistics
In this paper, sentiment analysis is applied to one social media called Twitter. Sentiment analysis is categorized as a classification problem that can be solved using one of machine learning methods, namely Neural Network. If machine learning is applied, it is necessary to rebuilt the model from scratch using new training data that requires manual labelling process. Hence, it is better to apply other learning besides machine learning, such as transfer learning. The simulation in this research yielded an accuracy of transfer learning using Neural Network which will be tested by N-grams (bigram and trigram) feature and one of feature selection method, namely Extra-Trees Classifier. The highest value of transfer learning accuracy is obtained when one hidden layer, 250 neurons on hidden layer, and tanh activation function are used. The use of feature selection method in simulation can also improve the transfer learning performance, so that the accuracy value is higher than the one that does not use feature selection method.