Opinion mining on Indonesian tourism TikTok video content using fasttext and multilayer long short-term memory

Social media analysis is a trending topic among researchers, especially in exploring public opinion. The evolution of web 2.0 technology is a strong reason for turning social media into a digital platform that can easily facilitate various user expressions and opinions through diverse content. Expressions and opinions that emerge through interaction between social media users have great potential to be studied and used in various contexts, including for the government, which aims to understand the thoughts of its citizens regarding newly implemented public policies. Recently, the Government of the Republic of Indonesia established five super-priority tourist destinations through the Ministry of Tourism and Creative Economy (Kemenparekraf) to increase foreign exchange through tourism. These destinations are Lake Toba, Labuan Bajo, Borobudur, Mandalika, and Likupang. However, the policies launched need to be analyzed to support the success of the launched policies. One way is to analyze public opinion to find out how many citizens will recommend the destination. TikTok, one of the most used social media platforms on the market, can be used to investigate public opinion about specific tourist locations. Nowadays, many young travelers use TikTok to express their thoughts and feelings. Due to the non-standard language and the frequent use of slang in daily interactions, extracting opinions on TikTok’s social media data presents specific difficulties. It might be beneficial to utilize a language corpus frequently used to analyze public sentiment more accurately. This study used FastText word embedding combined with the Long Short - Term Memory (LSTM) model with single, double, and triple layers to investigate the public’s opinion of Tiktok social media data. Based on the experiment, using FastText and LSTM with multiple layers provides good performance in developing various system innovations for investigating public opinion, especially on social media data TikTok.