Analisis Penerapan Machine Learning Dalam Prediksi Risiko Keterlambatan Proyek Konstruksi

Pambudi, Hengki Jayeng (2026) Analisis Penerapan Machine Learning Dalam Prediksi Risiko Keterlambatan Proyek Konstruksi. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032232036-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (849kB) | Request a copy

Abstract

Keterlambatan proyek konstruksi merupakan salah satu tantangan utama yang berdampak pada biaya, jadwal, dan manfaat ekonomi. Penelitian ini bertujuan untuk menganalisis penerapan algoritma Machine Learning dalam memprediksi risiko keterlambatan proyek konstruksi, khususnya pada proyek Refinery Development Master Plan (RDMP) Balikpapan. Data historis proyek periode 2020–2022 digunakan untuk membangun model prediksi dengan metode Support Vector Machine (SVM), Random Forest (RF), dan model hibrida Random Forest yang dioptimasi menggunakan Genetic Algorithm (RF-GA). Proses analisis meliputi pembersihan data, transformasi variabel, analisis deskriptif, pemodelan, optimasi parameter, dan evaluasi performa model menggunakan metrik Accuracy, Precision, Recall, F1-score, dan AUC. Hasil analisis deskriptif menunjukkan bahwa disiplin Construction menjadi penyumbang deviasi kumulatif terbesar, sedangkan disiplin Procurement dan Engineering relatif stabil. Variabel deviasi kumulatif (Cumulative Dev), deviasi periode berjalan (This Period Dev), dan deviasi sebelumnya (Previous Dev) terbukti menjadi faktor utama penyebab keterlambatan proyek. Evaluasi performa model menunjukkan bahwa RF-GA memiliki akurasi tertinggi sebesar 99,34%, dengan nilai Precision, Recall, dan F1-score masing-masing 0,99 serta AUC mencapai 1,000, mengungguli SVM dan RF tanpa optimasi. Model RF-GA juga mampu memprediksi risiko keterlambatan proyek pada periode 2024–2025 secara efektif, dengan dominasi keterlambatan pada disiplin Construction dan pola musiman yang konsisten. Temuan penelitian ini memiliki implikasi manajerial signifikan, yaitu pentingnya pemantauan deviasi waktu sejak tahap awal proyek, fokus pengendalian pada kegiatan lapangan, dan penerapan sistem prediktif berbasis data untuk mitigasi risiko keterlambatan. Secara ilmiah, penelitian ini membuktikan bahwa integrasi Machine Learning dengan optimasi Genetic Algorithm dapat meningkatkan akurasi prediksi, memberikan dasar pengambilan keputusan yang objektif, dan mendukung transformasi manajemen proyek menuju pendekatan berbasis data dan kecerdasan buatan (AI-assisted project management).
========================================================================================================================
Construction project delays are one of the main challenges that impact costs, schedules, and economic benefits. This study aims to analyze the application of Machine Learning algorithms in predicting the risk of Construction project delays, particularly in the Refinery Development Master Plan (RDMP) Balikpapan project. Historical project data from the period 2020–2022 was used to build predictive models using Support Vector Machine (SVM), Random Forest (RF), and a Hybrid Random Forest optimized with Genetic Algorithm (RF-GA). The analysis process includes Data cleaning, variable transformation, descriptive analysis, modeling, parameter optimization, and model performance evaluation using Accuracy, Precision, Recall, F1-score, and AUC metrics. Descriptive analysis results show that the Construction discipline contributes the most to Cumulative Deviation, while the Procurement and Engineering disciplines remain relatively stable. The variables Cumulative Deviation (Cumulative Dev), current period deviation (This Period Dev), and Previous Deviation (Previous Dev) are proven to be the main factors causing project delays. Model performance evaluation shows that RF-GA has the highest Accuracy of 99.34%, with Precision, Recall, and F1-score all at 0.99 and an AUC of 1.000, outperforming SVM and non-optimized RF. The RF-GA model is also able to predict project Delay risks for the period 2024–2025 effectively, with delays dominated in the Construction discipline and consistent seasonal patterns. The findings of this study have significant managerial implications, namely the importance of monitoring time deviations from the early project stages, focusing control on field activities, and implementing a data-driven predictive system for Delay risk mitigation. Scientifically, this study demonstrates that the integration of Machine Learning with Genetic Algorithm optimization can improve prediction Accuracy, provide a basis for objective decision-making, and support the transformation of project management towards a data-driven and AI-assisted project management approach.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Machine Learning, Random Forest, Genetic Algorithm, prediksi keterlambatan, proyek konstruksi, RDMP Balikpapan
Subjects:	T Technology > T Technology (General) > T56.8 Project Management
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Hengki Jayeng Pambudi
Date Deposited:	27 Jul 2026 08:49
Last Modified:	27 Jul 2026 08:49
URI:	http://repository.its.ac.id/id/eprint/139004

Actions (login required)

View Item