TY  - THES
PB  - Universitas Pakuan
N2  - Student Final Project Topic Classification System Using the
XGBoost Algorithm
Nabila Rasman Sutandi1
, Boldson Herdianto Situmorang2
, Dinar Munggaran Akhmad3*
1,2,3 Department of Computer Science, Faculty of Mathematics and Natural Science, Pakuan
University, Bogor, West Java, 16143, Indonesia
Abstract
The final project is one of the requirements that must be met by students to complete their education in
higher education. The data on the title of the final project of the students of the Computer Science Study
Program at Pakuan University is still being recorded based on the student graduation period and has not been
classified based on the topic of the final project, so that students who will carry out the final project have
difficulty finding references for the title of the final project that suits the topic. This research aims to create a
classification system for student final project topics using the Extreme Gradient Boosting (XGBoost)
algorithm. The development of the system is assisted by the Term Frequency-Inverse Document Frequency
(TF-IDF) method and is classified into 4 topics, namely Artificial Intelligence, Hardware Programming,
Computer Networking, and Software Engineering. This study used 1079 final project title data from 2018-
2022, which was divided into 3 comparisons of training data and test data, namely 70:30 (model 1), 80:20
(model 2), and 90:10 (model 3). Parameter tuning is carried out using gridsearchCV and k-fold cross validation
to get the best parameters. The results showed that model 3 had the best performance with an accuracy of
87.85%. The XGBoost model in the system automatically predicts the title of the final project along with its
topic label, so admins don't need to add topics manually. Users (students) can view and search for final project
title data based on topics that have been classified to be used as a reference for new final project titles according
to the topic.
Keywords: Final Project Topic; Classification; System; XGBoost; TF-IDF
AV  - public
M1  - Skripsi
Y1  - 2024/08/09/
ID  - eprintsunpak8882
UR  - http://eprints.unpak.ac.id/8882/
A1  - Sutandi, Nabila Rasman
A1  - Herdianto Situmorang, Boldson
A1  - Munggaran Akmad, Dinar
TI  - Student Final Project Topic Classification System Using the
XGBoost Algorithm
EP  - 27
ER  -