Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FLEXIBLE: Forecasting Cellular Traffic by Leveraging Explicit Inductive Graph-Based Learning

Duc Thinh Ngo, Kandaraj Piamrat, Ons Aouedi, Thomas Hassan, Philippe Raipin-Parvédy

TL;DR

Flexible addresses cellular traffic forecasting under evolving network topologies by learning inductively from local k-hop subgraphs. The model combines a dilated causal temporal convolutional network with a Gin-like graph aggregation to capture both temporal dynamics and local spatial correlations, and it supports unseen base stations and straightforward transfer learning across cities. It achieves up to 9.8% improvement over state-of-the-art baselines, with pronounced advantages when training data are scarce, demonstrating practical applicability for new enbs and continual network growth.

Abstract

From a telecommunication standpoint, the surge in users and services challenges next-generation networks with escalating traffic demands and limited resources. Accurate traffic prediction can offer network operators valuable insights into network conditions and suggest optimal allocation policies. Recently, spatio-temporal forecasting, employing Graph Neural Networks (GNNs), has emerged as a promising method for cellular traffic prediction. However, existing studies, inspired by road traffic forecasting formulations, overlook the dynamic deployment and removal of base stations, requiring the GNN-based forecaster to handle an evolving graph. This work introduces a novel inductive learning scheme and a generalizable GNN-based forecasting model that can process diverse graphs of cellular traffic with one-time training. We also demonstrate that this model can be easily leveraged by transfer learning with minimal effort, making it applicable to different areas. Experimental results show up to 9.8% performance improvement compared to the state-of-the-art, especially in rare-data settings with training data reduced to below 20%.

FLEXIBLE: Forecasting Cellular Traffic by Leveraging Explicit Inductive Graph-Based Learning

TL;DR

Flexible addresses cellular traffic forecasting under evolving network topologies by learning inductively from local k-hop subgraphs. The model combines a dilated causal temporal convolutional network with a Gin-like graph aggregation to capture both temporal dynamics and local spatial correlations, and it supports unseen base stations and straightforward transfer learning across cities. It achieves up to 9.8% improvement over state-of-the-art baselines, with pronounced advantages when training data are scarce, demonstrating practical applicability for new enbs and continual network growth.

Abstract

From a telecommunication standpoint, the surge in users and services challenges next-generation networks with escalating traffic demands and limited resources. Accurate traffic prediction can offer network operators valuable insights into network conditions and suggest optimal allocation policies. Recently, spatio-temporal forecasting, employing Graph Neural Networks (GNNs), has emerged as a promising method for cellular traffic prediction. However, existing studies, inspired by road traffic forecasting formulations, overlook the dynamic deployment and removal of base stations, requiring the GNN-based forecaster to handle an evolving graph. This work introduces a novel inductive learning scheme and a generalizable GNN-based forecasting model that can process diverse graphs of cellular traffic with one-time training. We also demonstrate that this model can be easily leveraged by transfer learning with minimal effort, making it applicable to different areas. Experimental results show up to 9.8% performance improvement compared to the state-of-the-art, especially in rare-data settings with training data reduced to below 20%.
Paper Structure (32 sections, 8 equations, 2 figures, 5 tables)

This paper contains 32 sections, 8 equations, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related works
  3. Graph Neural Networks - GNNs
  4. Spatiotemporal forecasting
  5. Transfer learning
  6. Methodology
  7. Problem definition
  8. Graph construction
  9. Proximity graph
  10. k-hop subgraphs
  11. Spatiotemporal module
  12. Temporal Convolutional Network
  13. Graph Isomorphism Network
  14. ReadIn and ReadOut
  15. ReadIn
  16. ...and 17 more sections

Figures (2)

  • Figure 1: The proposed frameworks: (a) transferring Flexbile from source to target city and (b) model architecture which takes a batch of k-hop subgraphs and associated historical traffic to generate a batch of prediction for the target enb.
  • Figure 2: The MAE at 15min-horizon for the amount of data used for training.