Sentiment Analysis on Dirty Vote Movie in X (Twitter) Using Convolutional Neural Network

This research focuses on Sentiment Analysis of Dirty Vote based on Twitter users' opinions by utilizing the Convolutional Neural Network (CNN) method. In this study, a total of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 0 1 0}$</tex> tweets were used after going through the labeling and pre-processing process, with a division of training data and test data of 80:20. The results showed a True Positive value of 44, True Negative of 104, False Positive of 26, and False Negative of 28. The accuracy value obtained was 73.27 % with a precision of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 0 \%}$</tex>, recall of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{7 8. 8 \%}$</tex>, specificity of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{6 2. 8 6 \%}$</tex>, and f1-score of 79.4%. Word cloud analysis revealed that the majority of opinions on the Dirty Vote movie tend to be positive, indicating significant support or appreciation from the public. Based on these results, future research is recommended to expand the data collection period, integrate more data sources from Other algorithms from machine learning and deep learning fields were used, such as KNN, SVM, Naïve Bayes, and LSTM. Improvements in the pre-processing stage are also recommended to increase the accuracy of the model, especially in overcoming the use of slang and slang terms that are often used on social media.