Klasifikasi Emosi Teks pada Media Sosial Platform X Menggunakan Metode K-Nearest Neighbor

Abstract

According to the Kamus Besar Bahasa Indonesia (KBBI), emotion is defined as a surge of feelings that fluctuates rapidly, as well as psychological and physiological states and reactions such as joy, sadness, melancholy, affection, and courage which are subjective in nature. In expressing their emotions, humans can articulate them verbally or through writing. With technological advancements, research has been conducted on human-computer emotion interaction, particularly emotions extracted from text. Emotion analysis in text can be applied across various media, one of which is microblogging sites like Platform X. With the rapid growth of data, there is a need to classify data to quickly generate structured information, even in large volumes, without consuming significant time. All data to be categorized are divided into five emotion categories: Anger, Fear, Enthusiasm, Sadness, and Happiness. This study utilizes the K-Nearest Neighbor method and 2000 tweets. The K- nearest neigbor the results of k-fold cross validation testing show an accuracy of 88%. based on the research findings, the K-nearest neighbor method successfully classifies emotions on Twitter.