Identifikasi Gejala Depresi Pada Komentar Media Sosial Menggunakan Metode Gated Recurrent Unit (GRU)

Abstract

The growth of social media is very swift, and it was recorded that as much as 61.8% of the opulation in Indonesia uses the social media. Social media is used for differentpurposes, such as to find instructions, opinions, and suport,even solution to a problemthe user is currently facing, one of them is comments that shows symptoms of depression. The prevalence of cases of depression in Indonesia itself reaches 3.7% ofthe total population, and the suicide rate reaches close to 4 times the amount reported.Lack of knowledge about depression could potentially worsen the user’s condition if not treated right. Therefore, an approach to know symptoms of depression in socialmedia comments is necessary, to avoid replies that will worsen the condition ofthe user who is expressing what they are experiencing. This study conducts sentimentanalysis with Gated Recurrent Unit (GRU) on 10000 data of social media comments.The accuracy achieved by the model in this study is 93.2%.