DualClusterNet: Personalized Pattern-Space Composition for Next-POI Recommendation
Organizations: Shanghai Jiao Tong University Shanghai, China · East China Normal University Shanghai, China
Abstract
Next point-of-interest (POI) recommendation supports personalized place discovery on location-based Web platforms. Mobility routines can recur across users while being realized at different places and adapted to individual preferences. We present DualClusterNet, a causal sequential recommender that composes two population-shared prototype dictionaries. At each observed prefix, soft routers retrieve pattern and spatial mixtures; a factorized multiplicative interaction couples them, feature-wise affine modulation adapts the shared representation, and a separate user-conditioned branch with an additive--projective gate retains individual variation. The prediction head combines this representation with a learned first-order transition term and ranks the complete training POI vocabulary. We evaluate on NYC and TKY using a training-only vocabulary and an original-final-target protocol that neither bridges unknown POIs nor substitutes earlier targets. The available single-seed results show competitive full-candidate ranking, while controlled analyses support the contribution of the compositional and personalized branches. Code is available at https://anonymous.4open.science/r/DualClusterNet-F4EE/.
Figures & tables
| Item | NYC | TKY |
|---|---|---|
| Training users | 1,047 | 2,262 |
| Training POI candidates | 4,980 | 7,793 |
| Training categories | 318 | 289 |
| Training sequences | 11,022 | 35,166 |
| Validation queries | 1,355 | 4,770 |
| Test queries | 1,300 | 4,711 |
| NYC | TKY | |||||||
|---|---|---|---|---|---|---|---|---|
| Method | Acc@1 | Acc@5 | Acc@10 | MRR | Acc@1 | Acc@5 | Acc@10 | MRR |
| FPMC ( Rendle et al., 2010 ) | 0.1700 | 0.4369 | 0.5438 | 0.2917 | 0.0951 | 0.2585 | 0.3600 | 0.1790 |
| ST-RNN ( Liu et al., 2016 ) | 0.2054 | 0.4862 | 0.5923 | 0.3352 | 0.2309 | 0.4698 | 0.5636 | 0.3408 |
| STGN ( Zhao et al., 2019 ) | 0.2115 | 0.4838 | 0.5877 | 0.3361 | 0.2437 | 0.4787 | 0.5602 | 0.3520 |
| STGCN ( Zhao et al., 2019 ) | 0.2162 | 0.4777 | 0.5662 | 0.3334 | 0.2428 | 0.4751 | 0.5621 | 0.3499 |
| PLSPL ( Wu et al., 2022 ) | 0.2269 | 0.4777 | 0.5446 | 0.3361 | 0.2371 | 0.4458 | 0.5256 | 0.3365 |
| Variant | NYC | TKY | ||
|---|---|---|---|---|
| Acc@1 | MRR | Acc@1 | MRR | |
| Full | 0.3007 | 0.4209 | 0.3403 | 0.4507 |
| Backbone | 0.2713 | 0.3926 | 0.3147 | 0.4263 |
| w/o pattern | 0.2837 | 0.4058 | 0.3269 | 0.4376 |
| w/o spatio | 0.2864 | 0.4083 | 0.3291 | 0.4405 |
| Concat pattern–space | 0.2926 | 0.4147 | 0.3338 | 0.4449 |
| Sweep | Setting | NYC | TKY |
|---|---|---|---|
| Pattern | 8 | 0.2726 | 0.3147 |
| 16 | 0.2821 | 0.3228 | |
| 0.2873 | 0.3261 | ||
| 64 | 0.2891 | 0.3274 | |
| 128 | 0.2856 | 0.3239 | |
| Spatial | 8 | 0.2764 | 0.3169 |
| Case | Path | Full | Back. | Outcome | |
|---|---|---|---|---|---|
| C1 | A–B–C D | 1 | 7 | 0.7213 | Full only |
| C2 | E–F–E G | 1 | 1 | 0.4387 | Both correct |
| C3 | H–I–J K | 9 | 1 | 0.3129 | Backbone only |
| C4 | L–M–N O | 18 | 12 | 0.5816 | Both fail |
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
| Method | Retained mechanism | Model-visible inputs | Local adaptation and training objective |
|---|---|---|---|
| FPMC | User–candidate preference plus first-order previous–next POI factorization | User and observed POI prefix | Singleton next-POI form; BPR with 10 uniformly sampled negatives per valid training step |
| ST-RNN | Recurrent transition with linearly interpolated temporal and distance matrices | User, POI, observed elapsed-time and distance gaps | Log-spaced interpolation over 12 time and 12 distance bins; full-vocabulary cross-entropy (CE) |
| STGN | Separate spatio-temporal recurrent gates | User, POI, observed time and distance gaps | Uses only observed gaps and a user residual at scoring; full CE |
| STGCN | Parameter-reduced coupled-gate variant introduced with STGN | Same as STGN | This name denotes the coupled STGN cell, not a graph convolution; full CE |
| PLSPL | POI-short, category-short, and causal long-term branches with a user-specific mixture | User, POI, category, hour, and weekday | Causal prefix attention; no future category or time; full CE |
| STAN | Two attention stages with learned spatial/temporal interval interpolation | User, POI, hour, weekday, training coordinates, and observed elapsed time | Uses the latest observed time, recomputes each prefix causally, and chunks only computation (not candidates); sampled CE with 10 negatives |
| Method | Singleton architecture setting | Parameters (NYC/TKY) | Best epoch (NYC/TKY) |
|---|---|---|---|
| FPMC | MF dimension 64; Markov dimension 64; dropout 0 | 1,028,212 / 1,648,881 | 19 / 7 |
| ST-RNN | hidden 64; 12 time and 12 distance bins; maxima 720 h and 500 km; dropout 0.2 | 811,956 / 1,252,593 | 30 / 29 |
| STGN | POI embedding 64; hidden 64; dropout 0.2 | 759,540 / 1,200,177 | 25 / 25 |
| STGCN | POI embedding 64; hidden 64; coupled gates; dropout 0.2 | 751,284 / 1,191,921 | 14 / 27 |
| PLSPL | POI embedding 64; context 32; hidden 96; dropout 0.2 | 1,057,433 / 1,551,923 | 28 / 28 |
| STAN | embedding 50; candidate-computation chunk 512; prefix checkpointing enabled | 317,850 / 519,250 | 17 / 24 |
Sequential Recommendation
Sequential recommendation predicts what a user will choose next from ordered interactions, adapting to changing interests and context.
See all 153 papers