Pith. sign in

REVIEW 4 major objections 5 minor 37 references

GCN-TULHOR: Trajectory-User Linking Leveraging GCNs and Higher-Order Spatial Representations

T0 review · 4 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read GCN-TULHOR claims that adding a single graph-convolutional layer to a hex-tessellation transformer consistently improves trajectory-user linking, by 1–8% relative accuracy and F1, across sparse check-in and continuous GPS datasets.

desk verdict A reasonable incremental extension of the authors' own TULHOR, but the evidence is thinner than the abstract claims, and an unstated train/test graph split could be leaking test information into the GCN. read the letter →

arxiv 2509.11095 v1 pith:SY4YRW4V submitted 2025-09-14 cs.LG

Khoa Tran , Pranav Gupta , Manos Papagelis This is my paper · ORCID
classification cs.LG
keywords trajectory-userlinkinggraphconvolutionalnetworkhexagonaltessellationhigher-ordermobilityrepresentationspatialembeddingstrajectoryclassificationdataTULHOR
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper seeks to establish that a graph-based spatial reasoning layer materially improves the task of matching anonymized mobility traces to their generating users. It builds on TULHOR, which converts raw trajectories into sequences of hexagonal cells, and adds a graph convolutional network that learns spatial embeddings from a data-driven adjacency graph of those cells. Across six real-world datasets—sparse check-ins and continuous GPS—the augmented model reports consistent gains over TULHOR and other sequence-only baselines, with relative improvements of 1–8% in accuracy and F1. If the gains hold, they suggest that explicitly modeling spatial dependencies as graph structure is a complementary and transferable signal for trajectory-user linking.

What carries the argument

The key machinery is the GCN layer with a data-driven adjacency matrix over hexagonal grid cells. Each cell is a node; edges are defined by the union of observed consecutive transitions in trajectories (weighted by frequency) and geometrically adjacent cells not observed (weighted 1 as a smoothing prior). The matrix is row-normalized, then symmetrically normalized with self-loops, and the GCN aggregates neighbor embeddings to produce spatial context vectors that feed into a non-invasive self-attention transformer. This graph construction carries the argument because it lets the model learn spatial dependencies adaptively from the data itself, rather than from a fixed topology.

What would settle it

Re-run GCN-TULHOR and TULHOR with multiple random seeds on the Tokyo check-in dataset (|U|=209) and check whether the F1 difference (0.875 vs 0.876) exceeds run-to-run variance; also inspect the graph-construction code to confirm that test trajectories never contribute edges to the adjacency matrix.

Watch

Extended reading notes

Core claim

The central discovery is that a single GCN layer, operating on a graph whose nodes are hexagonal cells and whose edges are derived from observed trajectory transitions plus geometric adjacency, improves trajectory-user linking beyond what sequence models alone achieve. The adjacency matrix weights observed transitions by their frequency and fills gaps with a structural prior of geometric neighborliness, so the GCN captures both data-driven and topological spatial relationships without needing timestamps or POIs. On the reported experiments, GCN-TULHOR outperforms TULHOR and other baselines on all six datasets, with improvements most pronounced on continuous trajectories (e.g., HO-Geolife Acc

Load-bearing premise

The reported improvements are assumed to be stable signal rather than run-to-run noise, and the GCN adjacency graph is assumed to be constructed without any information from test trajectories; if either fails, the 1–8% gains would not survive scrutiny.

Editorial extensions

If this is right

  • Adding a GCN spatial-embedding layer to a transformer-based trajectory model yields measurable accuracy improvements without side information such as timestamps or POI categories.
  • The optimal configuration of a single GCN layer suggests that one message-passing step is sufficient to capture useful spatial locality; deeper layers hurt by over-smoothing node distinctions.
  • GCN-learned spatial embeddings transfer to weaker sequence models: an LSTM augmented with GCN embeddings substantially outperforms a plain LSTM on the Rome dataset, indicating the spatial knowledge is reusable.
  • The unified hexagonal higher-order representation makes the method applicable to both sparse check-in and continuous GPS data with the same architecture, improving generalizability across mobility modalities.
  • The data-driven adjacency graph—combining observed transition frequencies with geometric adjacency—is the mechanism that lets the model capture both common routes and unobserved but spatially near regions.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the reported gains are stable, a natural extension is that richer graph constructions (e.g., edge attributes encoding time-of-day or travel speed) could yield further improvements; the paper does not explore such edge features.
  • A reader should verify the graph-construction code: the paper does not explicitly state whether the adjacency matrix is built only from training trajectories. If test trajectories contribute edges, the gains would be inflated by label leakage.
  • The sharp drop in performance as GCN depth increases hints that for trajectory-user linking, a single round of neighborhood aggregation is enough; deeper aggregation may blur individual identity signals that are already well captured by hex-cell abstraction.
  • The transfer result to LSTM suggests that GCN embeddings could serve as a general spatial pretraining feature for other mobility tasks such as next-location prediction or trip classification, not just user linking.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes GCN-TULHOR, an extension of the TULHOR framework for trajectory-user linking. Raw check-in and GPS trajectories are mapped to hexagonal-grid 'higher-order mobility flow' sequences, and a graph convolutional network is used to produce spatial embeddings over a graph whose nodes are hex cells and whose edges combine observed trajectory transitions with geometric adjacency. These spatial embeddings are integrated with TULHOR's transformer-style encoder via non-invasive self-attention. The authors report experiments on check-in and continuous trajectory datasets, claiming 1–8% relative gains in accuracy and F1 over TULHOR and other baselines, with sensitivity analyses on GCN depth, embedding size, and hex resolution. Source code is promised in a public repository.

Significance. If the claims held, the contribution would be a simple and reusable recipe: adding one GCN layer over a data-driven hex-cell adjacency graph improves trajectory-user linking on both sparse and continuous data, with transferable spatial embeddings. The paper has useful ingredients: a unified hex-based representation for heterogeneous trajectory data, a clearly specified graph construction combining observed transitions with a geometric prior, and a transferability experiment (GCN embeddings improving an LSTM). The source-code availability is a strength. However, the current evidence is not sufficient to support the central claim: several reported improvements are at or below the level of run-to-run noise, no uncertainty quantification is provided, and one table contains internally inconsistent copy-paste-looking values. Most importantly, the paper does not state whether the adjacency graph is built from training trajectories only, leaving open a test-data leakage pathway that would invalidate the comparison.

major comments (4)
  1. [§4.4.5 and §5.1] The graph-construction description is ambiguous with respect to the train/test split. §4.4.5 defines N_obs(i) as hexagons that 'directly follow v_i in any trajectory', and states the adjacency matrix is 'computed dynamically from the trajectory dataset itself'. §5.1 describes an 80/20 split for check-in data but never states that graph construction is restricted to the training split. If all trajectories, including test trajectories, contribute transitions to N_obs, then each test cell's GCN embedding aggregates co-occurrence statistics with other cells in the same test trajectories, giving the GCN model access to test-set pairwise information that the sequence-only TULHOR baseline does not have. This would make the reported gains an artifact of leakage. The authors must specify and implement train-only adjacency construction, and report results under that condition; the current manuscri
  2. [Tables 4 and 5] The abstract claims 'consistent improvements' and '1–8% relative gains in accuracy and F1'. The check-in results do not support this. In Table 4, the |U|=108 row shows precision changing by -2.17% (0.938 to 0.945? no, improvement is -2.17% relative to the strongest baseline, but the GCN-TULHOR precision is still numerically higher than TULHOR; however the 'Improvement' row is against the strongest baseline, not TULHOR). More importantly, Table 5 shows many gains below 0.3%: e.g., |U|=209 F1 goes 0.875 to 0.876, |U|=451 Acc@1 goes 0.801 to 0.802 and precision is flat at 0.783. Negative deltas appear for precision/recall in some rows. No error bars, multiple seeds, or significance tests are reported, so these differences are indistinguishable from noise. The paper's claim of consistent, meaningful gains is not established by the reported evidence.
  3. [Table 7] The HO-TDrive rows in Table 7 are internally inconsistent and appear to be duplicated from the HO-Porto rows in Table 6. TULHOR is listed as 0.090/0.074/0.070/0.090/0.070 and GCN-TULHOR as 0.100/0.160/0.080/0.100/0.080 for both HO-Porto (Table 6) and HO-TDrive (Table 7). The reported 'Improvement' for HO-TDrive is 1.04% for Acc@1, but the table values imply 11.1%; Acc@5 implies 116%, not 8.67%. This is a data integrity problem: either the HO-TDrive results are wrong, or the table was copied incorrectly. The findings for T-Drive cannot be assessed until the correct results are provided.
  4. [§5.4] The prose overstates the reported effect. For the Foursquare-TKY dataset, the text says 'meaningful and consistent improvements', but the numbers in Table 5 are almost all within 0.1–0.3% relative (e.g., Acc@1 from 0.893 to 0.894 at |U|=209, F1 from 0.875 to 0.876), and some metrics decrease. The manuscript's own Limitations section admits that 'it can sometimes achieve only marginal improvements, performing on par with baseline methods in certain scenarios.' The evaluation text should be calibrated to match the actual effect sizes, or supported by repeated runs with confidence intervals.
minor comments (5)
  1. [§4] Unresolved placeholder '(add ref to related)' appears in the first paragraph of Section 4. This should be completed before submission.
  2. [Throughout] The dataset name is misspelled as 'Foursqare' in several places (e.g., Tables 4 and 5, Section 5.4). Please correct to 'Foursquare'.
  3. [§5.1 / Abstract] The abstract and introduction state experiments on 'six real-world datasets', but Section 5.1 and Tables 4–7 list eight variants (NYC-CI, TKY-CI, NYC-CON, TKY-CON, Porto, Rome, Geolife, T-Drive). Please clarify the count or the naming.
  4. [Figures 2 and 3] Figure 3 is referenced in the text before Figure 2, and the figure numbers appear out of order in the material (Figure 2 is shown after Figure 3). Reordering would improve readability.
  5. [Table 3] Table 3 is malformed: entries such as 'HO-NYC-CI|U| =108 6,489 108 6,489 3,197' do not clearly map to columns |T|, |U|, |T uniq|, |H uniq|, and the same numeric value appears in multiple columns. This makes the dataset statistics difficult to interpret.

Circularity Check

1 steps flagged · score 4.0 of 10

Potential circularity: GCN adjacency is defined over 'all observed trajectories' with no stated train/test restriction, so reported gains may be produced by test-set transitions.

  1. fitted input called prediction [Section 4.4.5 (Adjacency Matrix Definition) with Eq. 4 in Section 4.1 and the split in Section 5.1]
    "Let V be the set of all hexagonal cells in the spatial grid and T the set of all observed trajectories, each represented as an ordered list of visited hexagon indices. For each node v_i∈V, let N_obs(i)⊂V be the set of neighboring hexagons that directly follow v_i in any trajectory... the adjacency matrix in GCN-TULHOR is computed dynamically from the trajectory dataset itself."

    The GCN spatial embedding z_s_i = GCN(g_i, G) (Eq. 4) is computed from graph G whose edges and weights are defined by N_obs(i) over T (Eqs. 16-18). Section 5.1 splits the data into 80% training and 20% test sets, but never restricts graph construction to the training split. If T includes the test trajectories, each test cell's GCN embedding aggregates transition counts from the very trajectories being classified. The reported test-set improvements of GCN-TULHOR over the graph-free TULHOR baseline are then partly supplied by test-set statistics, so the evaluation is forced by construction rather than by learned spatial generalization. The paper's failure to specify a train/test boundary for the adjacency matrix leaves this circularity unresolved.

full rationale

Aside from the adjacency-matrix issue, the paper's central comparison is empirical rather than derivational: GCN-TULHOR is measured against external baselines and the authors' own TULHOR on held-out trajectories, and the reported gains are not produced by an algebraic identity. The use of TULHOR as the primary baseline is a legitimate ablation-style comparison, not load-bearing circularity, and no uniqueness theorem or ansatz is imported from the authors' prior work. However, the adjacency matrix definition in Section 4.4.5 is a data-driven input to the GCN, and the evaluation section does not state that this graph is built exclusively from the training split; if all trajectories contribute edges, the GCN sees test-set transition patterns that the sequence-only baseline cannot, making the claimed 1-8% gains an artifact. This is a concrete, quotable gap in the derivation chain, even though it may be resolved by the available code. The manuscript also contains an explicit missing-support placeholder ('limitations mentioned in (add ref to related)') and marginal-improvement acknowledgements, but those are completeness/transparency issues rather than circularity. Overall, the core method is not circular by definition, but the unguarded graph construction creates a partial circularity in the evaluation.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the spatial abstraction pipeline and the graph definition; no new physical entities are introduced. The main unstated burden is the leakage-free construction of the adjacency graph.

free parameters (4)
  • Hexagonal resolution level = HEX8 for most results; HEX9 for some continuous datasets
    Tuned via sensitivity analysis (Table 8); controls cell size, sparsity, and performance.
  • Number of GCN layers = 1
    Chosen as optimal in Section 5.6; more layers decrease F1.
  • Embedding dimension = 512
    Selected in sensitivity study; 1024 slightly worse.
  • Balanced loss beta = 0.99
    Set manually in Eq. (14) to weight classes.
assumptions (3)
  • domain assumption Routing algorithms (e.g., OSRM) infer plausible routes between consecutive check-ins
    Invoked in Section 3.4; if inferred routes are unrealistic, the higher-order flow representation and the GCN edges built from it are corrupted.
  • domain assumption Hexagonal tessellation preserves user-discriminative spatial information at the chosen resolution
    The whole pipeline depends on this; Table 8 shows performance varies strongly across HEX levels.
  • ad hoc to paper The adjacency matrix is built from training trajectories only
    Implied by the 80/20 split but never stated; if test transitions are included, the GCN sees test structure during training.

how reviews work

0 comments
Cite this review

Pith. "Pith review of GCN-TULHOR: Trajectory-User Linking Leveraging GCNs and Higher-Order Spatial Representations." pith.science (2026). https://pith.science/paper/SY4YRW4V

@misc{pith2026250911095,
  author       = {Pith},
  title        = {Pith review of: GCN-TULHOR: Trajectory-User Linking Leveraging GCNs and Higher-Order Spatial Representations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SY4YRW4V}},
  note         = {Machine review of arXiv:2509.11095}
}
read the original abstract

Trajectory-user linking (TUL) aims to associate anonymized trajectories with the users who generated them, which is crucial for personalized recommendations, privacy-preserving analytics, and secure location-based services. Existing methods struggle with sparse data, incomplete routes, and limited modeling of complex spatial dependencies, often relying on low-level check-in data or ignoring spatial patterns. In this paper, we introduced GCN-TULHOR, a method that transforms raw location data into higher-order mobility flow representations using hexagonal tessellation, reducing data sparsity and capturing richer spatial semantics, and integrating Graph Convolutional Networks (GCNs). Our approach converts both sparse check-in and continuous GPS trajectory data into unified higher-order flow representations, mitigating sparsity while capturing deeper semantic information. The GCN layer explicitly models complex spatial relationships and non-local dependencies without requiring side information such as timestamps or points of interest. Experiments on six real-world datasets show consistent improvements over classical baselines, RNN- and Transformer-based models, and the TULHOR method in accuracy, precision, recall, and F1-score. GCN-TULHOR achieves 1-8% relative gains in accuracy and F1. Sensitivity analysis identifies an optimal setup with a single GCN layer and 512-dimensional embeddings. The integration of GCNs enhances spatial learning and improves generalizability across mobility data. This work highlights the value of combining graph-based spatial learning with sequential modeling, offering a robust and scalable solution for TUL with applications in recommendations, urban planning, and security.

Figures

Figures reproduced from arXiv: 2509.11095 by the authors.

Figure 1
Figure 1. An illustrative example showing how sample check-in data from [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. An example (Foursquare-NYC) showing the transformation from higher-order check-in data to higher-order mobility flow data, enriching trajectory semantics and helping address the TUL problem. Definition 6 (Higher-order Check-in). Each check-in location 𝑝 ∈ P is mapped to the grid cell 𝑔 ∈ G that contains it. Definition 7 (Higher-order Trajectory). A trajectory 𝑇 𝑟 = {𝑝1, 𝑝2, . . . , 𝑝𝑚} is transformed into a sequence… view at source ↗
Figure 2
Figure 2. Visualization of HO-Geolife trajectory data at different hexagonal resolutions. As resolution increases (from Hex@6 to Hex@10), [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Impact of higher-order abstraction on data sparsity for [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: An example (Foursquare-NYC-Continuous) showing how trajectory neighbors are leveraged in GCN-TULHOR 4.3 Pre-training and Fine-tuning GCN-TULHOR The model is pre-trained using a masked language modeling (MLM) objective: L𝑀𝐿𝑀 = 1 |𝑇 𝑟𝑚 | ∑︁ 𝑔𝑚 ∈𝑇 𝑟𝑚 − log 𝑃 (𝑔 𝑚 = 𝑔 𝑚∗ |…
Figure 6
Figure 6. Figure 6: High-level architecture of GCN-TULHOR. the connections between these locations. Each node is characterized by a feature vector derived from trajectory data, and the GCN learns node embeddings that encapsulate spatial dependencies.The GCN layer’s output embeddings enhan…
Figure 7
Figure 7. Figure 7: An illustrative example of how the adjacency matrix [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Impact of varying the number of GCN layers and embedding dimensions on the Macro-F1 score of the [PITH_FULL_IMAGE:figures/full_fig_p020_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 3 linked inside Pith

  1. [1]

    Mahmoud Alsaeed, Ameeta Agrawal, and Manos Papagelis. 2023. Trajectory-User Linking using Higher-order Mobility Flow Representations. In 2023 24th IEEE International Conference on Mobile Data Management (MDM) . 158–167. https://doi.org/10.1109/MDM58254.2023.00033

    arXiv 2023

  2. [2]

    Daniel Ashbrook and Thad Starner. 2003. Using GPS to Learn Significant Locations and Predict Movement Across Multiple Users. In Proceedings of the International Symposium on Wearable Computers . 101–108

    2003

  3. [3]

    Berndt and James Clifford

    Donald J. Berndt and James Clifford. 1994. Using Dynamic Time Warping to Find Patterns in Time Series. In Proceedings of the 3rd International Conference on Knowledge Discovery and Data Mining . 359–370

    1994

  4. [4]

    Lorenzo Bracciale, Paolo Loreti, and Sandro Mancini. 2022. CRA WDAD: A Publicly Available Dataset for Urban Mobility Research in Rome. In Proceedings of the 2022 Conference on Real-time and Intelligent Decision Support Systems . 22–31

    2022

  5. [5]

    Yuhan Chang, Ziyue Huang, Yunhao Zhang, Chao Zhang, Deren Lei, and Pan Li. 2025. HG-TUL: Heterogeneous Graph Neural Networks for Trajectory-User Linking. arXiv:2502.07549 [cs.LG] https://arxiv.org/abs/2502.07549

    arXiv 2025

  6. [6]

    Ling Chen and Raymond T. Ng. 2005. Robust and Fast Similarity Search for Moving Object Trajectories. In Proceedings of the IEEE International Conference on Data Engineering (ICDE) . 607–618

    2005

  7. [7]

    Thomas Eiter and Georg Gottlob. 1994. Computing Disjunctive Stable Models . Technical Report CD-TR 94/64. Institut für Informationssysteme, Technische Universität Wien. http://www.kr.tuwien.ac.at/staff/eiter/et-archive/files/cdtr9464.pdf Accessed May 2025

    1994

  8. [8]

    Soroush Faraji, Siamak Ravanbakhsh, Brian Karrer, and Aaron Clauset. 2023. Point2Hex: Efficient Hexagonal Tessellation for Spatial Data Representation. In Proceedings of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems . 78–87

    2023

Show all 37 references
  1. [9]

    Hao Gao, Xufu Liu, Meng Sun, Yuan Zhang, and Jun Wang. 2017. Identifying User Identity in Location-Based Social Networks with Recurrent Neural Networks. In Proceedings of the ACM Conference on Information and Knowledge Management (CIKM) . 1459–1468

    2017

  2. [10]

    Letian Gong, Yan Lin, Xinyue Zhang, Yiwen Lu, Xuedi Han, Yichen Liu, Shengnan Guo, Youfang Lin, and Huaiyu Wan. 2024. Mobility-LLM: Learning Visiting Intentions and Travel Preferences from Human Mobility Data with Large Language Models. In Proceedings of the 38th Conference on...

    2024

  3. [11]

    Yan Lin, Huaiyu Wan, Shengnan Guo, and Youfang Lin. 2022. Contrastive Pre-training of Spatial-Temporal Trajectory Embeddings. preprint arXiv:2207.14539 (2022)

    2022 arXiv

  4. [12]

    Chang Liu, Xiaoguang Li, Guohao Cai, Zhenhua Dong, Hong Zhu, and Lifeng Shang. 2021. Noninvasive self-attention for side information fusion in sequential recommendation. In Proc. of AAAI, Vol. 35. 4249–4256. Manuscript submitted to ACM 26 Tran, Gupta and Papagelis

    2021

  5. [13]

    Dennis Luxen and Christian Vetter. 2011. Real-time Routing with OpenStreetMap Data. In Proceedings of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems . 513–516

    2011

  6. [14]

    Wancan Miao, Xinna Wang, and Dongsheng Liu. 2020. DeepTUL: Deep Trajectory-User Linking with Attention Mechanism. In Proceedings of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems . 526–529

    2020

  7. [15]

    Luis Moreira-Matias, João Gama, Marcio Ferreira, Joana Mendes-Moreira, and Luis Damas. 2015. Taxi Trajectory Classification Using Hoeffding Trees. In Proceedings of the ACM Symposium on Applied Computing . 1674–1681

    2015

  8. [16]

    Paul Newson and John Krumm. 2009. Hidden Markov Map Matching Through Noise and Sparseness. In Proceedings of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems . 336–343

    2009

  9. [17]

    Lawrence R. Rabiner. 1989. A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition. Proc. IEEE 77, 2 (1989), 257–286

    1989

  10. [18]

    Shaojie Sun, Min Zhang, Xiaolin Liao, and Sheng Pan. 2018. Attentive Recurrent Neural Network for Trajectory Modeling and User Identification. In Proceedings of the IEEE International Conference on Data Mining (ICDM) . 317–326

    2018

  11. [19]

    Yifan Sun, Xu Ma, Tian Zhou, and Jun Ma. 2021. AttnTUL: Hierarchical Spatio-Temporal Attention for Trajectory-User Linking. In Proceedings of the IEEE International Conference on Data Mining (ICDM) . 641–650

    2021

  12. [20]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. In Advances in Neural Information Processing Systems 30 (NIPS 2017) . 5998–6008

    2017

  13. [21]

    Petar Veličković, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Liò, and Yoshua Bengio. 2018. Graph Attention Networks. In International Conference on Learning Representations (ICLR) . https://arxiv.org/abs/1710.10903

    2018 arXiv

  14. [22]

    Michalis Vlachos, Dimitrios Gunopulos, and Georgios Kollios. 2004. Discovering Similar Multidimensional Trajectories. IEEE Transactions on Knowledge and Data Engineering 16, 8 (2004), 1019–1031

    2004

  15. [23]

    Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. 2019. How Powerful Are Graph Neural Networks?. In International Conference on Learning Representations (ICLR)

    2019

  16. [24]

    Dingqi Yang, Xiaohua Zhu, Jiawei Han, Hong Cheng, and Philip S. Yu. 2015. User Activity Modeling in Location-Based Social Networks. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . 1–10

    2015

  17. [25]

    Di Yao, Gao Cong, Chao Zhang, and Jingping Bi. 2019. Computing trajectory similarity in linear time: A generic seed-guided neural metric learning approach. In ICDE. IEEE, 1358–1369

    2019

  18. [26]

    Di Yao, Chao Zhang, Jianwei Zhu, Qingshan Hu, and Zhihua Chen. 2017. Learning deep representation for trajectory clustering. In International Conference on Machine Learning and Cybernetics (ICMLC) , Vol. 2. 443–448

    2017

  19. [27]

    Di Yao, Chao Zhang, Jianwei Zhu, Qingshan Hu, Zhihua Chen, and Jian Wang. 2018. Trajectory representation learning for human mobility. In AAAI. 4118–4125

    2018

  20. [28]

    Guan Yuan, Penghui Sun, Jie Zhao, Daxing Li, and Canwei Wang. 2017. A review of moving object trajectory clustering algorithms. Artificial Intelligence Review (2017), 123–144

    2017

  21. [29]

    Jing Yuan, Yu Zheng, Xing Xie, and Guangzhong Sun. 2011. Driving with Knowledge from the Physical World. In Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining . ACM, 316–324

    2011

  22. [30]

    Yansu Yuan, Yu Zheng, Xing Xie, and Guangzhong Sun. 2010. T-Drive: Driving Directions Based on Taxi Trajectories. InProceedings of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems . 99–108

    2010

  23. [31]

    Liang Zhao, Fei Li, and Junfeng Zhu. 2022. Spatio-Temporal Attention Graph Neural Network for Traffic Forecasting. Electronics 13, 1 (2022), 212. https://www.mdpi.com/2079-9292/13/1/212

    2022

  24. [32]

    Yu Zheng. 2015. Trajectory data mining: an overview. ACM Transactions on Intelligent Systems and Technology (TIST) 6, 3 (2015), 1–41

    2015

  25. [33]

    Yu Zheng, Xing Xie, and Wei-Ying Ma. 2010. GeoLife: A Collaborative Social Networking Service among User, Location, and Trajectory. IEEE Data(base) Engineering Bulletin (2010)

    2010

  26. [34]

    Yu Zheng, Li Zhang, Xiang Xie, and Wen-Chih Ma. 2008. Understanding Mobility Based on GPS Data. In Proceedings of the International Conference on Ubiquitous Computing (UbiComp) . 312–321

    2008

  27. [35]

    Yu Zheng, Lei Zhang, Xing Xie, and Wei-Ying Ma. 2009. Mining Interesting Locations and Travel Sequences from GPS Trajectories. InProceedings of the 18th International World Wide Web Conference (WWW ’09) . ACM, 791–800

    2009

  28. [36]

    Fan Zhou, Qiang Gao, Goce Trajcevski, Kunpeng Zhang, Ting Zhong, and Fengli Zhang. 2018. Trajectory-User Linking via Variational AutoEncoder.. In IJCAI. 3212–3218

    2018

  29. [37]

    Xiaoyan Zhou, Zhe Li, He Yang, Chen Yang, and Yi Sun. 2021. GNNTUL: Graph Neural Network for Trajectory-User Linking. In Proceedings of the ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems . 45–54. Received 13 May 2025 Manuscript submitted to ACM

    2021

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.