Deep Learning Model for Identification of Indonesian National Figure Entities on Social Media Using LSTM Architecture

This study aims to develop a Long Short-Term Memory (LSTM)-based Deep Learning model to identify Indonesian national figures in Indonesian social media texts. The background of this research is based on the increasing role of social media as a primary source of public information, which often discusses national figures such as officials, artists, and activists. However, the characteristics of informal language, the use of abbreviations, and non-standard spellings on social media pose challenges for traditional rule-based or statistical Named Entity Recognition (NER) systems. This study utilizes 1,109 tweets from Twitter obtained through the Twitter API, covering the names of popular public figures in various fields. The research process includes data crawling, preprocessing, labeling using the spaCy library, dividing training and test data, and training an LSTM model. The model architecture consists of an embedding layer, a bidirectional LSTM, and a time-distributed dense layer with a softmax activation function. The evaluation results show excellent model performance, with an accuracy increase of up to 97.8%, precision of 96%, recall of 93%, and F1-score of 92% on the validation data. Word cloud analysis shows that the model is able to consistently recognize Person entities such as "Prabowo", "Sri Mulyani", and "Agnez Mo". However, the model still experiences limitations in detecting unfamiliar or rarely appearing entities. Overall, this study shows that the combination of spaCy and LSTM is effective for NER tasks on Indonesian social media texts and has the potential for further development with increased data variety and improvements to the labeling process.

Keywords: Named Entity Recognition, LSTM, social media, national figures, NLP