Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{197432,
        author = {Ms. Sadia Kausar and Md Zuber Ali and Abdul Rahman Anas and Amaan Ahmed and Ahmed Waliuddin Quadri},
        title = {Adaptive Feature Fusion Networks for Origin-Destination Passenger Flow Prediction in Metro Systems},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {11},
        pages = {6274-6286},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=197432},
        abstract = {Accurate prediction of Origin-Destination (OD) passenger flow is a critical component of intelligent metro transportation systems, enabling efficient scheduling, congestion mitigation, and improved passenger experience. Unlike traditional station-level inflow and outflow prediction, OD prediction provides a fine-grained understanding of passenger movement across the entire metro network. However, this task is inherently challenging due to complex spatial dependencies between stations, highly dynamic temporal patterns, and the presence of sparse and incomplete OD matrices caused by delayed trip completion.
In this paper, we propose an advanced deep learning framework termed Adaptive Feature Fusion Network (AFFN) to address these challenges. The proposed model integrates heterogeneous data sources and captures multi-dimensional dependencies through a unified architecture. Specifically, an Enhanced Multi-Graph Convolution Gated Recurrent Unit (EMGC-GRU) is designed to model spatial relationships by leveraging multiple knowledge-based graphs and automatically learned attention-based graphs to capture hidden correlations among stations. To incorporate temporal dynamics, the GRU structure is embedded with graph convolution operations, enabling simultaneous learning of spatial-temporal patterns.},
        keywords = {},
        month = {April},
        }