Pengembangan Aplikasi Android Prediksi Kesehatan Mental Mahasiswa menggunakan Algoritma XGBoost

Mental health among students currently needs to be taken very seriously. Therefore, there is a need for an appropriate solution that can help students address this issue. This research aims to develop an application that is expected to help students perform self-diagnosis with a higher level of accuracy compared to self-diagnosis based solely on literature obtained from the internet. In this research, several development stages will be carried out. The stages include data collection and analysis, machine learning model development, model deployment to API, and Android application development. In the machine learning model development stage, the process of creating a machine learning model to predict students' mental health by classifying it into three classes low, medium, and high indications was conducted. Based on the evaluation results, the XGBoost model is the most optimal with an accuracy rate of 90.4%, making it the main model chosen for this research. Meanwhile, in the application development stage, based on the System Usability Scale evaluation results from 35 respondents, an average score of 73.8 was obtained, which falls into the Acceptable category with a Good rating. Thus, the developed application is deemed feasible and can be used by students to predict mental health independently with a better accuracy level.