Analisis Sentimen Media Sosial Twitter Pada Kasus Tragedi Kanjuruhan Dengan Menggunakan Long Short-Term Memory (LSTM)

Twitter is not only a platform for interaction and discussion, but also reflects a diversity of topics, including profound issues such as the tragedy at Kanjuruhan Stadium in Malang City. This tragedy, which occurred on October 1, 2022, shocked the Indonesian football world by claiming the lives of 135 people. The response from the football community, including players, clubs, and supporters, created a wave of moral support and demands to thoroughly investigate this incident. A lively discussion took place on Twitter, where supporters voiced their opinions, creating significant pros and cons. This research aims to analyze public sentiment regarding the Kanjuruhan tragedy using the Long Short-Term Memory (LSTM) method in sentiment analysis. LSTM, as a form of deep learning, is used to process text data and measure sentiment tendencies, whether positive, negative, or neutral. With a sample size of 5,061, the analysis results show that the LSTM model achieved an Accuracy rate of 94%, Positive Precision of 92% and Negative Precision of 97%, Positive Recall of 97% and Negative Recall of 90%, Positive F1-Score of 94%, and Negative F1-Score of 93%. The conclusion of this research provides an overview of the dominance of negative sentiment in response to the Kanjuruhan tragedy on Twitter, while the LSTM model proves its effectiveness in analyzing public sentiment. As a future suggestion, the research could be expanded with the addition of training data and special attention to the pre-processing stage to improve the quality of the analysis.