1. Combination of convolutional neural network and long short-term memory to enhance the sentiment analysis result with the Indonesian language.
- Author
-
Yusli, D., Maylawati, D. S., Ramdhani, M. A., Jumadi, J., and Zulfikar, W. B.
- Subjects
- *
CONVOLUTIONAL neural networks , *SENTIMENT analysis , *INDONESIAN language , *NATURAL language processing , *MATHEMATICAL convolutions - Abstract
Sentiment analysis is computational research of opinion and emotion that covers textually. This research aims to investigate the Natural Language Processing (NLP) technology in analyzing the sentiment from social media with the Indonesian language. From many methods applied in sentiment analysis, the popular one is Deep Learning (DL). Therefore, this study was conducted to increase sentiment analysis accuracy by knowing the effectiveness of the combination DL algorithm, among others the Convolutional Neural Network (CNN) algorithm and the Long Short Time Memory (LSTM) algorithm. Combining the CNN algorithm and the LSTM algorithm was carried out to process sentiment analysis with tweet data from Twitter. The pre-processing and weighting phase was done using Word2Vec. The result of the experiment with 1295 tweet data showed that the average accuracy of the CNN-LSTM is 71.19% with the highest accuracy of 75.04%. This result indicated that the CNN-LSTM algorithm has the highest accuracy results compared to the basic CNN algorithm and LSTM algorithm. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF