This is a list of useful information about urban mobility prediction. Related papers, datasets and codes are included.
| Papers | Authors | Code | Year | Venue | Performance |
|---|---|---|---|---|---|
| A recurrent model with spatial and temporal contexts.(ST-RNN)(Paper) | Qiang Liu(CAS), Shu Wu, Liang Wang, Tieniu Tan | Code | 2016 | AAAI | Gowalla: Rec@5 =0.1524, Rec@10=0.2714. GTD: Rec@5=0.4986, Rec@10=0.6812. |
| Geo-teaser: Geo-temporal sequential embedding rank for point-of-interest recommendation.(Geo-teaser)(Paper) | Shenglin Zhao(CUHK), Tong Zhao, Irwin King, and Michael R. Lyu | Code | 2017 | WWW | Foursquare: Prec@5=0.13, Prec@10=0.1, Rec@5=0.15, Rec@10=0.2 Gowalla: Prec@5=0.16, Prec@10=0.13, Rec@5=0.07, Rec@10=0.1 |
| Next point-of-interest recommendation with temporal and multi-level context attention.(TMCA)(Paper) | Ranzhen Li(SJTU), Yanyan Shen, Yanmin Zhu | Code | 2018 | ICDM | Gowalla: Rec@5=0.21926, Rec@10=0.27725. Foursquare: Rec@5=0.02870, Rec@10=0.04809. |
| HST-LSTM: A Hierarchical Spatial-Temporal Long-Short Term Memory Network for Location Prediction.(HST-LSTM)(Paper) | Dejiang Kong(ZJU), Fei Wu | None | 2018 | IJCAI | Baidu Map: Acc@10=0.4847, Acc@20=0.5657. |
| Content-aware hierarchical point-of-interest embedding model for successive poi recommendation.(CAPE)(Paper) | Buru Chang(KU), Yonggyu Park, Donghyeon Park, Seongsoon Kim, Jaewoo Kang | Code | 2018 | IJCAI | With STELLAR: Rec@5=0.2384, Rec@10=0.2989. With LSTM: Rec@5=0.2412, Rec@10=0.3054. With GRU: Rec@5=0.2433, Rec@10=0.3079. With ST-RNN: Rec@5=0.2239, Rec@10=0.2601. |
| DeepMove: Predicting Human Mobility with Attentional Recurrent Networks.(DeepMove)(Paper) | Jie Feng(THU), Yong Li, Chao Zhang, Funing Sun, Fanchao Meng, Ang Guo, Depeng Jin | Code | 2018 | WWW | Foursquare (NY): Rec@5=0.3372, Rec@10=0.4091. Gowalla: Rec@5=0.2021, Rec@10=0.2510. |
| Long-and short-term preference learning for next poi recommendation.(LSPL)(Paper) | Yuxia Wu(XJTU), Ke Li, Guoshuai Zhao, Xueming Qian | None | 2019 | CIKM | Foursquare (NYC): Prec3@10=0.3901, Prec@20=0.4461. Foursquare (TKY): Prec@10=0.3986, Prec@20=0.4596. |
| Where to go next: A spatio-temporal gated network for next poi recommendation.(STGN)(Paper19, Paper20) | Pengpeng Zhao(SUDA), Haifeng Zhu, Yanchi Liu, Jiajie Xu, Zhixu Li, Fuzhen Zhuang, Victor S. Sheng, Xiaofang Zhou | None | 2019 | AAAI (2020 TKDE) | Foursquare (CA): Acc@5=0.1308, Acc@10=0.1612. Foursquare (SIN): Acc@5=0.2737, Acc@10=0.3017. Gowalla: Acc@5=0.1644, Acc@10=0.2020. Brightkite: Acc@5=0.4953, Acc@10=0.5231. |
| Topic-Enhanced Memory Networks for Personalised Point-of-Interest Recommendation.(TEMN)(Paper) | Xiao Zhou(Cambridge), Cecilia Mascolo, Zhongxiang Zhao | Code | 2019 | KDD | WeChat (GPR): TEMN (GPR): Acc@5=0.70389, Acc@10=0.81752. TEMN (CPR): Acc@5=0.72876, Acc@10=0.83398. |
| Location prediction over sparse user mobility traces using rnns: Flashback in hidden states!(Flashback)(Paper) | Dingqi Yang(UM), Benjamin Fankhauser, Paolo Rosso, Philippe Cudre-Mauroux | Code | 2020 | IJCAI | Foursquare: Acc2@5=0.5399, Acc@10=0.6236. Gowalla: Acc@5=0.2754, Acc@10=0.3479. |
| Discovering subsequence patterns for next poi recommendation!(ASPPA)(Paper) | Kangzhi Zhao(THU) | None | 2020 | IJCAI | Foursquare (US): Acc@10=0.3371, Acc@20=0.3950. Gowalla: Acc@10=0.2947, Acc@20=0.3573. |
| Next Point-of-Interest Recommendation on Resource-Constrained Mobile Devices.(LLRec)(Paper) | Qinyong Wang(UQ) | None | 2020 | WWW | Foursquare: Acc@10=0.3542, Acc@20=0.4594. Gowalla: Acc@10=0.3874, Acc@20=0.4781. |
| Personalized Long- and Short-term Preference Learning for Next POI Recommendation.(PLSPL)(Paper) | Yuxia Wu(XJTU), Ke Li, Guoshuai Zhao, Xueming Qian | None | 2020 | TKDE | Foursquare (NYC): Prec@10=0.3953, Prec@20=0.4475 Foursquare (TKY): Prec@10=0.4020, Prec@20=0.4664. |
| Where to go next: Modeling long-and short-term user preferences for point-ofinterest recommendation.(LSTPM)(Paper) | Ke Sun(WHU), Tieyun Qian, Tong Chen, Yile Liang, Quoc Viet Hung Nguyen, Hongzhi Yin | Code | 2020 | AAAI | Foursquare (NY): Rec@5=0.3372, Rec@10=0.4091. Gowalla: Rec@5=0.2021, Rec@10=0.2510. |
| An Interactive Multi-Task Learning Framework for Next POI Recommendation with Uncertain Check-ins.(iMTL)(Paper) | Lu Zhang( NTU) | None | 2020 | IJCAI | Foursquare (CLT): Rec@10=0.0534, Map4@10=0.0238. Foursquare (CAL): Rec@10=0.0691, Map@10=0.0443. Foursquare (PHO): Rec@10=0.0769, Map@10=0.0352. |
| A Category-Aware Deep Model for Successive POI Recommendation on Sparse Check-in Data.(CatDM)(Paper) | Fuqiang Yu(SDU), Lizhen Cui, Wei Guo, Xudong Lu, Qingzhong Li, Hua Lu | Code | 2020 | WWW | Foursquare (NYC): Rec@5=0.2407, Rec@10=0.3113. Foursquare (TKY): Rec@5=0.2148, Rec@10=0.2739. |
| An attentional recurrent neural network for personalized next location recommendation.(ARNN)(Paper) | Qing Guo(NTU), Zhu Sun, Jie Zhang, Yin-Leng Theng | None | 2020 | WWW | Foursquare (NY): Acc@10=0.4162, Acc@20=0.4393 Foursquare (TK): Acc@10=0.4285, Acc@20=0.4864 Gowalla (SF): Acc@10=0.2336, Acc@20=0.2530. |
| Exploiting geographical-temporal awareness attention for next point-of-interest recommendation.(GT-HAN)(Paper) | Tongcun Liu(BUPT), Jianxin Liao, Zhigen Wu, Yulong Wang, Jingyu Wang | None | 2020 | Neurocomputing | Foursquare: AUC8=0.9661, acc@5: 0.13-0.15, acc@10: 0.17-0.19, acc@20: 0.23-0.25 (depending on latent dimensionality). |
| Time-Aware Location Prediction by Convolutional Area-of-Interest Modeling and Memory-Augmented Attentive LSTM.(t-LocPred)(Paper) | Chi Harold Liu(BIT), Yu Wang, Chengzhe Piao, Zipeng Dai, Ye Yuan, Guoren Wang, Dapeng Wu | None | 2020 | TKDE | Gowalla: MRR5=0.247 (C=6, all), Weeplaces: MRR=0.277 (C=6, all), Brightkite: MRR=0.388 (C=4, all). |
| Content-Aware Successive Point-of-Interest Recommendation.(CAPRE)(Paper) | Buru Chang(KU), Yookyung Koh, Donghyeon Park, Jaewoo Kang | None | 2020 | SDM | Foursquare: Rec@5=0.1724, Rec@10=0.2084 Instagram: Rec@5=0.2934, Rec@10=0.3588. |
| Geography-Aware Sequential Location Recommendation.(GeoSAN)(Paper) | Defu Lian(USTC), Yongji Wu, Yong Ge, Xing Xie, Enhong Chen | Code | 2020 | KDD | Foursquare: Acc@5=0.3735, Acc@10=0.4867. Gowalla: Acc@5=0.4951, Acc@10=0.6028. Brightkite: Acc@5=0.5258, Acc@10=0.6425. |
| Modeling hierarchical category transition for next POI recommendation with uncertain check-ins.(HCT)(Paper) | Lu Zhang(NTU), Zhu Sun, Jie Zhang, Horst Kloeden, Felix Klanner | None | 2020 | Information Sciences, Elsevier | Foursquare(SIN): Prec@5=0.613 Rec@5=0.0403 Foursquare(NYC): Prec@5=0.0585, Rec@5=0.0352 Foursquare(LA): Prec@5=0.0653, Rec@5=0.0305. |
| HME: A Hyperbolic Metric Embedding Approach for Next-POI Recommendation.(HME)(Paper) | Shanshan Feng(Abu Dhabi, UAE), Lucas Vinh Tran, Gao Cong, Lisi Chen, Jing Li, Fan Li | None | 2020 | SIGIR | Foursquare (NYC): Rec@5=0.0962, Rec@10=0.1371. Foursquare (TKY): Rec@5=0.1527, Rec@10=0.2172. Gowalla (Houston): Rec@5=0.1533, Rec@10=0.2318. |
| Location Prediction via Bi-direction Speculation and Dual-level Association.(paper) | Xixi Li(WHU),Ruimin Hu, Zheng Wang,Toshihiko Yamasaki | None | 2021 | ijcai | Gowalla: Acc@1=0.1454, Acc@5=0.3531, Acc@10=0.4192, MRR=0.2431. Foursquare: Acc@1=0.3068, Acc@5=0.6612, Acc@10=0.7136, MRR=0.4505. |
| SNPR A Serendipity-Oriented Next POI Recommendation Model.(SNPR)(paper) | Mingwei Zhang(Northeastern University Shenyang, China),Yang Yang,Rizwan Abbas,Ke Deng,Jianxin Li,Bin Zhang | None | 2021 | CIKM | Foursquare(NewYork): Prec@3=0.05695, Prec@5=0.05325, Prec@10=0.04829, Prec@20=0.03425. Foursquare(United Kingdom): Prec@3=0.04346, Prec@5=0.03842, Prec@10=0.03651, Prec@20=0.02502. |
| LightMove: A Lightweight Next-POI Recommendation forTaxicab Rooftop Advertising.(LightMove)(paper) | Jinsung Jeon(Yonsei University,Seoul, South Korea),Soyoung Kang,Minju Jo, Seunghyeon Cho,Noseong Park,Seonghoon Kim,Chiyoung Song | code | 2021 | CIKM | Texi: Hits@1=0.9988, Hits@5=1.0000, Hits@10=1.0000, MRR=0.9994. Foursquare: Hits@1=0.1545, Hits@5=0.3203, Hits@10=0.3656, MRR=0.2288. LA: Hits@1=0.3209, Hits@5=0.4431, Hits@10=0.4758, MRR=0.0.3756. |
| ST-PIL: Spatial-Temporal Periodic Interest Learning for Next Point-of-Interest Recommendation.(ST-PIL) (paper) | Qiang Cui(Meituan, Beijing, China), Chenrui Zhang, Yafeng Zhang, Jinpeng Wang, Mingchen Cai | None | 2021 | CIKM | NYC: Acc@1=0.3807, Acc@5=0.5850, Acc@10=0.6454, MRR@5=0.4584, MRR@10=0.4666. TKY: Acc@1=0.3523, Acc@5=0.5963, Acc@10=0.6772, MRR@5=0.4455， MRR@10=0.4564. |
| STAN: Spatio-Temporal Attention Network for Next Location Recommendation.(STAN)(paper) | Yingtao Luo(University of Washington),Qiang Liu,Zhaocheng Liu | code | 2021 | www | Gowalla: Recall@5 =0.3016, Recall@10=0.3998. TKY: Recall@5 =0.3461, Recall@10=0.4264. SIN: Recall@5 =0.3751, Recall@10=0.4301. NYC: Recall@5 =0.4669, Recall@10=0.5962. |
| Predicting Destinations by a Deep Learning based Approach(LATL)(paper) | Jiajie Xu(Soochow University),Jing Zhao,Rui Zhou,Chengfei Liu,Pengpeng Zhao,Lei Zhao | None | 2021 | TKDE | Beijing trajectory datasets: Acc@1=0.3570, Acc@5=0.6444, Acc@10=0.7165, Acc@20=0.8001. Chengdu trajectory datasets: Acc@1=0.3354, Acc@5=0.5344, Acc@10=0.6251, Acc@20=0.7163. |
| PREMERE: Meta-Reweighting via Self-Ensembling for Point-of-Interest Recommendation(PREMERE)(paper) | Minseok Kim(KAIST), Hwanjun Song, Doyoung Kim, Kijung Shin, Jae-Gil Lee | code | 2021 | AAAI | Gowalla: Precision@5=0.1389. Foursquare: Precision@5=0.0911. Yelp: Precision@5=0.0545. |
| Discovering Collaborative Signals for Next POI Recommendation with Iterative Seq2Graph Augmentation(SGRec)(paper) | Yang Li(The University of Queensland) , Tong Chen , Yadan Luo , Hongzhi Yin , Zi Huang | None | 2021 | ijcai | Foursquare: HR@1=0.195, HR@5=0.362, HR@10=0.402, HR@20=0.465. Gowalla: HR@1=0.067, HR@5=0.118, HR@10=0.137, HR@20=0.151. |
| MFNP: A Meta-optimized Model for Few-shot Next POI Recommendation(MFNP)(paper) | Huimin Sun(Soochow University),Jiajie Xu,Kai Zheng, Pengpeng Zhao,Pingfu Chao,Xiaofang Zhou | None | 2021 | ijcai | Foursquare: Acc@1=0.0255, Acc@5=0.0379, Acc@10=0.0503. Gowalla: Acc@1=0.0152, Acc@5=0.0308, Acc@10=0.0381. |
| Graph-Flashback Network for Next Location Recommendation(paper) | Xuan Rao(University of Electronic Science and Technology of China),Lisi Chen,Yong Liu,Shuo Shang, Bin Yao, Peng Han | code | 2022 | KDD | Gowalla: Acc@1=0.1512, Acc@5=0.3425, Acc@10=0.4256, MRR=0.2422. Foursquare: Acc@1=0.2805, Acc@5=0.5757, Acc@10=0.6514, MRR=0.4136. |
| MetaPTP: An Adaptive Meta-optimized Model for Personalized Spatial Trajectory Prediction(MetaPTP)(paper) | Yuan Xu(Soochow University),Jiajie Xu,Jing Zhao,Kai Zheng,An Liu,Lei Zhao | None | 2022 | KDD | Taxi trajectories in Beijing: Acc@1=0.6186, Acc@2=0.7864, Acc@3=0.8578, MRR=0.7467. Taxi trajectories in Porto: Acc@1=0.5310, Acc@2=0.7233, Acc@3=0.8176, MRR=0.6832. |
| Modeling Spatio-temporal Neighbourhood for Personalized Point-of-interest Recommendation(STGCAN)(paper) | Xiaolin Wang(Donghua University),Guohao Sun,Xiu Fang,Jian Yang, Shoujin Wang | code | 2022 | ijcai | NYC: Recall@1=0.257, Recall@5=0.544, Recall@10=0.629. TKY: Recall@1=0.171, Recall@5=0.384, Recall@10=0.457. Gowalla: Recall@1=0.129, Recall@5=0.343, Recall@10=0.414. |
| Next Point-of-Interest Recommendation with Inferring Multi-step Future Preferences(CFPRec)(paper) | Lu Zhang(Nanyang Technological University, Singapore),Zhu Sun,Ziqing Wu,Jie Zhang,Yew Soon Ong,Xinghua Qu | code | 2022 | ijcai | SIN: HR@5=0.2310, HR@10=0.3085, NDCG@5=0.1588, NDCG@10=0.1836. NYC: HR@5=0.2771, HR@10=0.3606, NDCG@5=0.1971, NDCG@10=0.2190. PHO: HR@5=0.3421, HR@10=0.4253, NDCG@5=0.2432, NDCG@10=0.2730. |
| Where to Go Next: A Spatio-Temporal Gated Network for Next POI Recommendation(NeuNext)(https://ieeexplore.ieee.org/document/9133505) | Pengpeng Zhao( Soochow University),Anjing Luo,Yanchi Liu,Jiajie Xu,Zhixu Li,Fuzhen Zhuang | None | 2022 | TKDE | Foursquare(CA): Acc@1=0.0890, Acc@5=0.1421, Acc@10=0.1748, MAP=0.2736. Foursquare(SIN): Acc@1=0.2284, Acc@5=0.2768, Acc@10=0.3022, MAP=0.3704. Gowalla: Acc@1=0.0930, Acc@5=0.1689, Acc@10=0.2034, MAP=0.2685. Brightkite: Acc@1=0.4567, Acc@5=0.5109, Acc@10=0.5422, MAP=0.5683. |
| Personalized Long- and Short-term Preference Learning for Next POI Recommendation(PLSPL)(paper) | Yuxia Wu(Xi’an Jiaotong University),KeLI,Guoshuai Zhao,Xueming Qian | None | 2022 | TKDE | Foursquare(NYC): Recall@1=0.1559, Recall@5=0.3252, Recall@10=0.3953, Recall@20=0.4475, MAP@5=0.2172, MAP@10=0.2266, MAP@20=0.2302. Foursquare(TKY): Recall@1=0.1571, Recall@5=0.3321, Recall@10=0.4020, Recall@20=0.4664, MAP@5=0.2212, MAP@10=0.2307, MAP@20=0.2352. |
| Time-Aware Location Prediction by Convolutional Area-of-Interest Modeling and Memory-Augmented Attentive LSTM(t-LocPred)(paper) | Chi Harold Liu(Beijing Institute of Technology),Yu Wang,Chengzhe Piao,Zipeng Dai, Ye Yuan,Guoren Wang, and Dapeng Wu | None | 2022 | TKDE | Weeplaces: Acc@1=0.277. Gowalla: Acc@1=0.247. Brightkite: Acc@1=0.380. |
| SPATM: A Social Period-Aware T opic Model for Personalized Venue Recommendation(SPATM)(paper) | Weiyu Ji(Beijing University of Posts and Tele-communications),Xiangwu Meng,Yujie Zhang | None | 2022 | TKDE | Foursquare: Recall@1=0.0683, Recall@5=0.157, Recall@10=0.222, Recall@15=0.273, NDCG@1=0.0683, NDCG@5=0.0374, NDCG@10=0.0297, NDCG@15=0.0255. Yelp: Recall@1=0.0243, Recall@5=0.0918, Recall@10=0.161, Recall@15=0.2432, NDCG@1=0.0243, NDCG@5=0.0196, NDCG@10=0.0173, NDCG@15=0.0168. |
| Learning Graph-based Disentangled Representations for Next POI Recommendation(DRAN)(paper) | Zhaobo Wang(Shanghai Jiao Tong University),Yanmin Zhu∗,Haobing Liu,Chunyang Wang | None | 2022 | sigir | Foursquare: Recall@2=0.3551, Recall@5=0.4092, Recall@10=0.4512. Gowalla: Recall@2=0.2288, Recall@5=0.2832, Recall@10=0.3291. |
| GETNext: Trajectory Flow Map Enhanced Transformer for Next POI Recommendation(GETNext)(paper) | Song Yang(The University of Auckland),Jiamou Liu,Kaiqi Zhao | code | 2022 | sigir | FourSquare(NYC): Acc@1=0.2435, Acc@5=0.5089, Acc@10=0.6143, Acc@20=0.6880, MRR=0.3621. FourSquare(TKY): Acc@1=0.2254, Acc@5=0.4417, Acc@10=0.5287, Acc@20=0.5829, MRR=0.3262. Gowalla(CA): Acc@1=0.1357, Acc@5=0.2852, Acc@10=0.3590, Acc@20=0.4241, MRR=0.2103. |
| Empowering Next POI Recommendation with Multi-Relational Modeling(MEMO)(paper) | Zheng Huang(University of Virginia), Jing Ma,Yushun Dong,Natasha Zhang Foutz,Jundong Li | None | 2022 | sigir | Baltimore July: Recall@10=0.891, MRR@10=0.446. DC July: Recall@10=0.831, MRR@10=0.380. DC August: Recall@10=0.840, MRR@10=0.435. |
| Next Point-of-Interest Recommendation with Auto-Correlation Enhanced Multi-Modal Transformer Network(paper) | Yanjun Qin(Beijing University of Posts and Telecommunications),Yuchen Fang,Haiyong Luo, Fang Zhao,Chenxing Wang | None | 2022 | sigir | NYC: Acc@5=0.4370, Acc@10=0.5289, MRR@5=0.2742, MRR@10=0.2867. TKY: Acc@5=0.3802, Acc@10=0.4687, MRR@5=0.2434, MRR@10=0.2553 |
| Predicting Human Mobility via Graph Convolutional Dual-attentive Networks(GCDAN)(paper) | Weizhen Dang(Tsinghua University),Haibo Wang, Shirui Pan,Pei Zhang, Chuan Zhou,Xin Chen,Jilong Wang | code | 2022 | wsdm | Gowalla: Acc@1=0.1377, Acc@5=0.3086, Acc@10=0.3780. Foursquare: Acc@1=0.1613, Acc@5=0.3417, Acc@10=0.4093. WiFi-Trace: Acc@1=0.5912, Acc@5=0.8064, Acc@10=0.8726. |
| RLMob: Deep Reinforcement Learning for Successive Mobility Prediction(RLMob)(paper) | Ziyan Luo(McGill University),Congcong Miao | code | 2022 | wsdm | Foursquare(TKY): Acc@1=0.4150. Foursquare(NYK): Acc@1=0.4401. Univ-WIFI: Acc@1=0.2291. |
| Hierarchical Multi-Task Graph Recurrent Network for Next POI Recommendation(paper) | Nicholas Lim(GrabTaxi Holdings),Bryan Hooi,See-Kiong Ng,Yong Liang Goh,Renrong Weng,Rui Tan. | code | 2022 | sigir | Gowalla: Acc@1=0.1455, Acc@5=0.2783, Acc@10=0.3394, Acc@20=0.4033. Foursquare: Acc@1=0.1673, Acc@5=0.3357, Acc@10=0.4148, Acc@20=0.4983. |