MACHINE LEARNING-BASED ANALYSIS AND RECOMMENDATIONS FOR ICONNET INFRASTRUCTURE DEVELOPMENT AREAS (CASE STUDY : BANDUNG REGENCY)

• Main Author: Novitri Susanti S.P., Rizki
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/86877
• Institution: Institut Teknologi Bandung

ICONNET is a Fiber to the home (FTTH) internet service provided by PLN Icon Plus, currently focusing on expanding its network distribution evenly across the West Java region. Bandung Regency, a densely populated area in this province, faces significant challenges in meeting the increasing demand for digital infrastructure. The high population density further complicates the issue of infrastructure distribution. The transmission medium used, fiber optik, despite its high investment cost, offers significant advantages such as large bandwidth capacity and high-speed data transmission, making it a worthwhile investment. Therefore, a comprehensive area recommendation analysis is required to support infrastructure development investments and ensure profitability for the company. This study utilizes datasets from various sources, including the Central Bureau of Statistics (BPS), PT PLN (Persero), and PLN Icon Plus. This study employs the Cross-Industry Standard Process for Data Mining (CRISP-DM) framework and utilizes datasets from various sources, such as the Central Bureau of Statistics (BPS), PT PLN (Persero), and PLN Icon Plus. The diversity of datasets with varying features and dimensions poses a significant challenge, which is addressed by using dimensionality reduction techniques such as Variational Autoencoder (VAE) and Principal Component Analysis (PCA). The analysis incorporates clustering algorithms like K-Means and DBSCAN for regional segmentation, as well as K-Nearest Neighbor (KNN) for predicting bandwidth requirements.