IMPLEMENTASI LARGE LANGUAGE MODEL (LLM) PADA SISTEM ASISTEN PENGINGAT BELAJAR BERBASIS INTERNET OF THINGS (IOT)

Ikhsan, Muhammad Toha (2026) IMPLEMENTASI LARGE LANGUAGE MODEL (LLM) PADA SISTEM ASISTEN PENGINGAT BELAJAR BERBASIS INTERNET OF THINGS (IOT). S1 thesis, Universitas PGRI Madiun.

	Text Halaman Depan.pdf Download (1MB)
	Text Abstrak.pdf Download (186kB)
	Text BAB I.pdf Download (258kB)
	Text BAB II.pdf Download (687kB)
	Text BAB III.pdf Restricted to Repository staff only Download (221kB) | Request a copy
	Text BAB IV.pdf Restricted to Repository staff only Download (3MB) | Request a copy
	Text BAB V.pdf Download (185kB)
	Text Daftar Pustaka.pdf Download (324kB)
	Text Lampiran.pdf Restricted to Repository staff only Download (991kB) | Request a copy

Abstract

The rapid development of digital technology has made the learning process more accessible. however, it has also increased the potential for distractions caused by digital devices. This study aims to implement a Large Language Model (LLM) in an Internet of Things (IoT)-based Study Reminder Assistant System that generates responses based on Pomodoro phases and the surrounding learning environment. An ESP32-S3 microcontroller serves as the main controller of the IoT device, which is equipped with a BMP280 temperature sensor, a BH1750 light sensor, and an INMP441 sound sensor. The LLM was first fine-tuned on an Indonesian-language dataset comprising three response modes: Pomodoro, which generates responses based on the current Pomodoro phase; Interruption, which responds to unfavorable environmental conditions; and Recovery, which provides responses when environmental conditions improve after an interruption. The generated LLM responses are displayed on the IoT device's screen, accompanied by simple facial-expression animations and supporting sound effects. The results show that the system was successfully implemented, achieving a 100% success rate in black-box testing (70 out of 70 test cases) and a 93.42% success rate in LLM evaluation using cosine similarity (213 out of 228 test scenarios). These results demonstrate that implementing an LLM in an IoT-based Study Reminder Assistant System can generate relevant responses based on the current Pomodoro phase and the user's learning environment.

Item Type:	Thesis/Skripsi/Tugas Akhir (S1)
Kata Kunci:	Study Reminder Assistant, Large Language Model, Internet of Things, Pomodoro, ESP32, Fine-tuning
Subjects:	L Education > LB Theory and practice of education Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software
Divisions:	Fakultas Teknik > Teknik Informatika
Depositing User:	IKHSAN TOHA MUHAMMAD
Date Deposited:	03 Aug 2026 07:01
Last Modified:	03 Aug 2026 07:01
URI:	http://eprint.unipma.ac.id/id/eprint/8202

Actions (login required)

View Item