relation: http://repository.upi.edu/146181/ title: IMPLEMENTASI INDOBERT DAN LARGE LANGUAGE MODEL PADA CHATBOT ADAPTIF BERDASARKAN EMOSI PENGGUNA creator: Destira Lestari Saraswati, - creator: Asep Wahyudin, - creator: Yudi Ahmad Hambali, - subject: Q Science (General) subject: QA75 Electronic computers. Computer science description: Tingginya angka gangguan kesehatan mental di Indonesia, khususnya pada generasi Milenial dan Generasi Z, belum diimbangi dengan ketersediaan layanan dukungan psikologis yang mudah diakses, terjangkau, dan bebas stigma. Penelitian ini mengembangkan aplikasi kesehatan mental berbasis chatbot yang memanfaatkan model IndoBERT untuk klasifikasi emosi meliputi cinta, marah, sedih, bahagia, dan takut serta mengintegrasikan Large Language Model untuk menghasilkan respons yang empatik dan adaptif sesuai kondisi emosional pengguna. Tujuan penelitian adalah merancang aplikasi chatbot yang memungkinkan pengguna mengekspresikan emosi secara anonim, menguji efektivitas aplikasi dalam mendukung pengguna mengelola kesehatan mental, dan mengevaluasi relevansi rekomendasi konten kesehatan mental yang dihasilkan sistem. Pengembangan dilakukan dengan pendekatan prototyping iteratif berbasis Flutter dan Firebase. Evaluasi sistem menggunakan pengujian Blackbox untuk memverifikasi fungsionalitas, kuesioner System Usability Scale untuk menilai aspek kegunaan, dan User Experience Questionnaire untuk mengevaluasi pengalaman pengguna. Hasil penelitian menunjukkan bahwa sistem bekerja sesuai spesifikasi, memperoleh skor rata-rata System Usability Scale sebesar 83.66 (di atas standar kelayakan 68), serta menghasilkan nilai positif pada enam dimensi User Experience Questionnaire, yaitu attractiveness (2,07), perspicuity (2,1), efficiency (2,04), dependability (1,96), stimulation (2,09), dan novelty (2,01). Temuan ini menegaskan bahwa aplikasi tidak hanya memenuhi aspek fungsionalitas dan kegunaan, tetapi juga memberikan pengalaman pengguna yang positif, terutama dalam hal kenyamanan curhat secara anonim dan penerimaan respons adaptif dari sistem. Penelitian ini masih terbatas pada penerapan ilmu komputer, sehingga penelitian selanjutnya perlu ditinjau lebih lanjut dari perspektif keilmuan psikologi. Penelitian ini memberikan kontribusi dalam integrasi pemrosesan IndoBERT dan Large Language Model berbasis bahasa Indonesia untuk menghadirkan solusi digital yang empatik, adaptif, dan ramah perangkat bergerak, sehingga berpotensi menjadi bentuk dukungan awal yang aman, terjangkau, dan inklusif bagi masyarakat. The high prevalence of mental health disorders in Indonesia, particularly among Millennials and Generation Z, has not been matched by the availability of psychological support services that are accessible, affordable, and free from stigma. This study developed a mental health application based on a chatbot that utilizes the IndoBERT model for emotion classification including love, anger, sadness, happiness, and fear and integrates a Large Language Model to generate empathetic and adaptive responses to users’ emotional states. The objectives of this research are to design a chatbot application that allows users to express emotions anonymously, to examine its effectiveness in supporting users to manage mental health, and to evaluate the relevance of mental health content recommendations generated by the system. The development process followed an iterative prototyping approach using Flutter for the interface and Firebase for the backend. System evaluation employed Blackbox testing to verify functionality, the System Usability Scale questionnaire to assess usability, and the User Experience Questionnaire to evaluate user experience. The results show that the system functioned as specified, achieved an average System Usability Scale score of 83.66 (above the standard threshold of 68), and produced positive results across all six dimensions of the User Experience Questionnaire: attractiveness (2,07), perspicuity (2,1), efficiency (2,04), dependability (1,96), stimulation (2,09), and novelty (2,01). These findings confirm that the application not only meets functional and usability standards but also provides a positive user experience, particularly in terms of comfort in anonymous self-disclosure and the reception of adaptive responses from the system. This study is limited to the application of computer science, and therefore future research should further examine the system from a psychological perspective. This research contributes to the integration of Natural Language Processing and Large Language Models in the Indonesian language, resulting in a digital solution that is empathetic, adaptive, and mobile-friendly, with the potential to serve as an accessible, affordable, and inclusive form of early psychological support for society. date: 2025-12-18 type: Thesis type: NonPeerReviewed format: text language: id identifier: http://repository.upi.edu/146181/1/S_KOM_2100506_Title.pdf format: text language: id identifier: http://repository.upi.edu/146181/3/S_KOM_2100506_Chapter1.pdf format: text language: id identifier: http://repository.upi.edu/146181/4/S_KOM_2100506_Chapter2.pdf format: text language: id identifier: http://repository.upi.edu/146181/5/S_KOM_2100506_Chapter3.pdf format: text language: id identifier: http://repository.upi.edu/146181/6/S_KOM_2100506_Chapter4.pdf format: text language: id identifier: http://repository.upi.edu/146181/2/S_KOM_2100506_Chapter5.pdf format: text language: id identifier: http://repository.upi.edu/146181/7/S_KOM_2100506_Appendix.pdf identifier: Destira Lestari Saraswati, - and Asep Wahyudin, - and Yudi Ahmad Hambali, - (2025) IMPLEMENTASI INDOBERT DAN LARGE LANGUAGE MODEL PADA CHATBOT ADAPTIF BERDASARKAN EMOSI PENGGUNA. S1 thesis, Universitas Pendidikan Indonesia. relation: https://repository.upi.edu/