Next-Generation Network Planning: Machine Learning (ml) Meets Real-World Mobile Data

In this next-generation network (NGN) age, data drives us to a future where innovation meets efficiency. The data-driven decisions fulfil rising traffic demands, minimize cost and lead to smarter network infrastructure. The mobile network operators (MNOs) can reduce their cost per bit by locating the area with the highest traffic density, represented in megabytes (MB) per $\text{km}^{2}$. This study proposes a base station (BS) clustering framework based on unsupervised machine learning (ML) to locate the target area for new deployments known as the highest traffic cluster (HTC). It is studied that the appropriate coverage range of cluster (in kilometers) represented as Epsilon is a significant factor in density-based network clustering to minimize the cost per MB and achieve higher network utilization. We propose a novel density-based learning technique assisted by real data to determine the appropriate $\epsilon$ and to locate the HTC. We have also studied the correlation of anomalous BSs (ABs) in the data in the context of network planning. The algorithm, density-based network clustering (DNC), determines the ABs, identifies the HTC and the appropriate value of epsilon by satisfying the MNOs’ requirements on the highest traffic density MB text km 2 and the target deployment area in text km2. We use the k nearest neighbours (k-NN) as a benchmark to compare the appropriate value of epsilon and other performance parameters.