Fig 1.
Spatial and temporal dependency among road sensors.
(a) Sensors 1 & 6 or sensors 4 & 5 show similar patterns, whereas they differ from sensor 2. (b) A daily traffic pattern is observed on three consecutive days.
Table 1.
Summary of Notation Used in the Framework.
Fig 2.
Overall architecture of the proposed Decomposition-based Dynamic Graph Adaptation of Large Language Models for Spatiotemporal Traffic Forecasting (DG-LLM) framework.
Fig 3.
Dynamic Graph Learning Pipeline for Mode-Specific Adjacency Construction.
Fig 4.
Mode-wise data processing through spatiotemporal embedding with graph-aware LLM backbone.
Table 2.
Statistics of the traffic forecasting datasets used in this study.
Table 3.
Short-Term comparison on NYC-Taxi Dataset (Pick-up and Drop-off).
Table 4.
Short-Term comparison on CH-Bike Dataset (Pick-up and Drop-off).
Table 5.
Short-Term comparison on PeMS Dataset (PeMS04 and PeMS08).
Fig 5.
Statistical significance of DG-LLM improvements.
Heatmaps show percentage reduction in MAE and RMSE relative to baseline models (n = 5 seeds), along with corresponding significance levels. Colors denote significance: purple (p < 0.001), red (p < 0.01), orange (p < 0.05), and grey (not significant). Upward arrows (↑) indicate error reduction by DG-LLM.
Fig 6.
Mean Absolute Error (MAE) comparison of DG-LLM against baseline models across 12 prediction horizons for all datasets.
Shaded areas represent the 95% confidence interval.
Table 6.
Long-Term Comparison of NYC-Taxi Drop-off and CH-Bike Drop-off.
Table 7.
Ablation Study Results on NYC-Taxi and CH-Bike Drop-off Prediction.
Fig 7.
Relative MAE and RMSE impact of ablation studies on the full model.
Fig 8.
MAE and RMSE comparison of curriculum learning strategy on NYC-Taxi and CH-Bike Dataset.
Fig 9.
VMD level comparisons on the NYC-Taxi Drop-off dataset.
Fig 10.
MAE & RMSE Comparisons for Different Unfrozen Layers on the NYC-Taxi and CH-Bike Datasets.
Fig 11.
MAE and RMSE Comparison on Different Pretrained and Non-Pretrained Backbones.
Fig 12.
MAE & RMSE Comparisons for Different Missing Rates on the NYC-Taxi and CH-Bike Datasets.
Table 8.
Computational efficiency of the proposed framework across datasets. Parameter counts are in millions (M).
Fig 13.
Mode-dependent graphs learned from decomposed traffic signals.
(a) Low Frequency, (b) Mid Frequency, and (c) High Frequency.
Fig 14.
Traffic forecasting visualization:
(a) Short-term, (b) Long-term.
Fig 15.
(a) Taxi drop-off dataset, (b) CH-Bike drop-off dataset, (c) PeMS04 dataset.
Fig 16.
MAE and RMSE for zero-shot cross-dataset transfer performance across varying urban modalities and data scales.