EN 
IN
Journal article
Text Mining pada Sosial Media untuk Mendeteksi Emosi Pengguna Menggunakan Metode Support Vector Machine dan K-Nearest Neighbour
Dwi Ardiada Made Sudarma IDA AYU DWI GIRIANTARI
Volume : 18 Nomor : 1 Published : 2019, April
Jurnal Ilmiah Teknik Elektro
Abstrak
Abstract— Twitter social networking and microblog services that allow users to send and read text-based messages up to 140 characters, known as tweets. A text on a tweet does not only convey information from an information, but also contains information about human behavior including emotions. To detect emotions from text on twitter social media services with unstructured data, one needs to do text analysis by using Text Mining. In this study propose to conduct text mining research on Social Media to detect users' emotions. From the tests conducted using the Support Vector Machine and K-Nearest Neighbor method, it can produce an average precision value of 0.4564. The recall value is 0.502 and the accuracy value is 0.8104 while the KNearest Neighbor method has an average precision value of 0.3421. The recall value is 0.4595 and the accuracy value is 0.797. The results of testing with the SVM-KNN method have a significant increase in the value of the K-Nearest Neighbor method. SVM-KNN method also has a significant difference in computational time in doing two classifications rather than the K-Nearest Neighbor method which only does one classification.