AMGST: Experimental Results for Traffic Forecasting on Four Public Datasets

Experimental results from the AMGST model for traffic forecasting, shared by author Pei Shi on figshare. The dataset, last updated on 2026-06-04, contains results from tests on four public traffic datasets measuring speed and flow. The file is 9.5 KB in size and is available in XLS format.

Use Cases

• Benchmarking new traffic forecasting models based on the reported performance metrics.
• Analyzing the effectiveness of adaptive multi-graph convolution for capturing spatial correlations.
• Studying the impact of spatiotemporal attention mechanisms on modeling long-term temporal dependencies in traffic data.

Strengths

• Results are derived from tests on four distinct public traffic datasets, providing a basis for comparison.
• The dataset is shared under a permissive CC-BY-4.0 license, allowing for reuse and redistribution.

Limitations

• The dataset is very small at 9.5 KB, indicating limited scope, likely containing only summary results.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for detailed analysis.

Provenance

Source

figshare, author Pei Shi.

Collection Method

Likely contains aggregated experimental results from machine learning model evaluation.

Freshness

Last updated 2026-06-04 17:36:38; freshness should be verified.