├── Datasets ├── README.md ├── auxiliary.md └── human_mobility.md ├── README.md └── Tasks ├── README.md ├── crowd.md ├── generative.md └── next-location.md /Datasets/README.md: -------------------------------------------------------------------------------- 1 | # Datasets - Deep Learning 4 Human Mobility 2 | ### [Back to Table Of Contents](https://github.com/scikit-mobility/DeepLearning4HumanMobility) 3 | --- 4 | 5 | In this part of the repository, you can access the datasets that can be used to model human mobility. 6 | You can explore two different categories of datasets: the ones strictly related to human mobility and others that cannot be used to directly model mobility but are strictly related to it. 7 | Regarding the former, you can access the following categories 8 | 9 | - [Human Mobility](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md) 10 | - [Mobile Phone Records](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#mobile-phone) 11 | - [GPS Traces](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#gps) 12 | - [Social Media](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#social-media) 13 | - [Others](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#others) 14 | 15 | While the auxiliary datasets are orgnanized as follows 16 | 17 | - [Auxiliary](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md) 18 | - [Spatial Aggregations](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#spatial-aggregations) 19 | - [Traffic](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#traffic) 20 | - [Environment](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#environment) 21 | - [Census and Administrative](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#census-and-administrative) 22 | -------------------------------------------------------------------------------- /Datasets/auxiliary.md: -------------------------------------------------------------------------------- 1 | # Spatial Aggregations 2 | 3 | | **Name** | **Type** | **Area** | **Link** | 4 | |:------------------:|:--------:|:--------:|:----------------------------------------------:| 5 | | Google S2 Geometry | Library | - | [bit.ly/S2Geometry](https://bit.ly/S2Geometry) | 6 | | Uber H3 | Library | - | [bit.ly/UberH3Geom](https://bit.ly/UberH3Geom) | 7 | | Scikit-Mobility | Library | - | [github.com/scikit-mobility](https://github.com/scikit-mobility/) | 8 | 9 | # Traffic 10 | 11 | | **Name** | **Type** | **Area** | **Link** | 12 | |:---------------------:|:---------:|:-------------------:|:------------------------------------------------------:| 13 | | MATSim | Simulator | - | [bit.ly/MATSim](https://bit.ly/MATSim) | 14 | | SUMO | Simulator | - | [bit.ly/SUMO-2](https://bit.ly/SUMO-2) | 15 | | Turin Traffic | Dataset | Turin (Italy) | [bit.ly/TorinoTraffic](https://bit.ly/TorinoTraffic) | 16 | | Hamburg Traffic | Dataset | Hamburg (Germany) | [bit.ly/HamburgTraffic](https://bit.ly/HamburgTraffic) | 17 | | France Traffic | Dataset | France | [bit.ly/FranceTraffic](https://bit.ly/FranceTraffic) | 18 | | New York City Traffic | Dataset | New York City (USA) | [bit.ly/NYCTrafficData](https://bit.ly/NYCTrafficData) | 19 | 20 | # Census and Administrative 21 | 22 | | **Name** | **Type** | **Area** | **Link** | 23 | |:---------------------:|:--------:|:--------------:|:----------------------------------------------------:| 24 | | ISTAT | Data Hub | Italy | [bit.ly/ItalyStat](https://bit.ly/ItalyStats) | 25 | | DANE | Data Hub | Colombia | [bit.ly/DANEColombia](https://bit.ly/DANEColombia) | 26 | | USCB | Data Hub | USA | [bit.ly/USCensus-3](https://bit.ly/USCensus-3) | 27 | | USCB - Commuting Data | Dataset | USA | [bit.ly/FlowsUS](https://bit.ly/FlowsUS) | 28 | | ONS - Commuting Data | Dataset | USA | [bit.ly/ONSFlows](https://bit.ly/ONSFlows) | 29 | | EU CensusHub | Data Hub | European Union | [bit.ly/EUCensusUB](https://bit.ly/EUCensusUB) | 30 | | UN Census Hub | Data Hub | - | [bit.ly/UNCensusHub](https://bit.ly/UNCensusHub) | 31 | | OECD | Data Hub | - | [bit.ly/OECDCensusHub](https://bit.ly/OECDCensusHub) | 32 | | EU Open Data Portal | Data Hub | European Union | [bit.ly/EUDataPortal](https://bit.ly/EUDataPortal) | 33 | | EU Urban Atlas | Data Hub | European Union | [bit.ly/EUUrbanAtlas](https://bit.ly/EUUrbasnAtlas) | 34 | 35 | # Environment 36 | 37 | | **Name** | **Type** | **Area** | **Link** | 38 | |:-------------------:|:-----------:|:----------------:|:--------------------------------------------------:| 39 | | Leeds | Dataset | Leeds (UK) | [bit.ly/LeedsMeteo](https://bit.ly/LeedsMeteo) | 40 | | SwissMetNet | Dataset | Switzerland | [bit.ly/SwissMet](https://bit.ly/SwissMet) | 41 | | US Meteo | Dataset | USA | [bit.ly/USWeather](https://bit.ly/USWeather) | 42 | | Seattle Weather | Dataset | Seattle (USA) | [bit.ly/SeattleMeteo](https://bit.ly/SeattleMeteo) | 43 | | Austin Weather | Dataset | Austin (USA) | [bit.ly/AustinMeteo](https://bit.ly/AustinMeteo) | 44 | | Delhi Weather | Dataset | Delhi (India) | [bit.ly/DelhiMeteo](https://bit.ly/DelhiMeteo) | 45 | | Szeged Weather | Dataset | Szeged (Hungary) | [bit.ly/SzegedMeteo](https://bit.ly/SzegedMeteo) | 46 | | Madrid Weather | Dataset | Madrid (Spain) | [bit.ly/MadridMeteo](https://bit.ly/MadridMeteo) | 47 | | Weather Underground | Web Service | - | [bit.ly/WUnderground](https://bit.ly/WUnderground) | 48 | | WorldClim | Web Service | - | [www.worldclim.org](https://www.worldclim.org/data/index.html) | 49 | 50 | -------------------------------------------------------------------------------- /Datasets/human_mobility.md: -------------------------------------------------------------------------------- 1 | # GPS 2 | 3 | | **Name** | **#Subjects** | **Timespan** | **Area** | **Access** | **Link** | 4 | |:----------------------------:|:-------------:|:------------:|:------------------------------------:|:-----------------------------:|:--------------------------------------------------------:| 5 | | Geolife | 182 | 3 years | Asia | Available | [bit.ly/Geolife](https://bit.ly/Geolife) | 6 | | T-Drive | 10,000 | 1 week | Beijing (China) | Available | [bit.ly/T-Drive-Data](https://bit.ly/T-Drive-Data) | 7 | | STResNet | - | - | Beijing (China), New York City (USA) | Available | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | 8 | | Taxi San Francisco | 500 | 1 month | San Francisco (USA) | Upon Registration | [bit.ly/TaxiSF](https://bit.ly/TaxiSF) | 9 | | Taxi Rome | 320 | 1 month | Rome (Italy) | Upon Registration | [bit.ly/TaxiRome](https://bit.ly/TaxiRome) | 10 | | Taxi Porto | - | - | Porto (Portugal) | Upon Registration | [bit.ly/TaxiPorto](https://bit.ly/TaxiPorto) | 11 | | SoBigData Tuscany Vehicles | 150,000 | 1 month | Tuscany Area (Italy) | On-site Upon Project Approval | [bit.ly/GPSTuscany](https://bit.ly/GPSTuscany) | 12 | | SoBigData Calabria Vehicles | 115,000 | 1 months | Calabria Area (Italy) | On-site Upon Project Approval | [bit.ly/GPSCalabria](https://bit.ly/GPSCalabria) | 13 | | SoBigData Volunteers' Tracks | - | 4 months | Tuscany Area (Italy) | On-site Upon Project Approval | [bit.ly/VolunteersTrack](https://bit.ly/VolunteersTrack) | 14 | | MDC | 185 | - | Lausanne (Switzerland) | Available | [bit.ly/MDC-2](bit.ly/MDC-2) | 15 | | Rio Buses | 12,000 | 1 month | Rio de Janeiro (Brasil) | Available | [bit.ly/RioBusData](https://bit.ly/RioBusData) | 16 | | Vessels | - | 1 month | Mediterranean Sea | Available | []() | 17 | | Apple's Report | - | - | World | Available | [bit.ly/AppleMobility](https://bit.ly/AppleMobility) | 18 | | Google's Report | - | - | World | Available | [bit.ly/GoogleMobility](https://bit.ly/GoogleMobility2) | 19 | 20 | # Mobile Phone 21 | 22 | | **Name** | **#Subjects** | **Timespan** | **Area** | **Access** | **Link** | 23 | |:--------------------------------:|:----------:|:---------:|:-------------------------------------:|:-----------------------------:|:---------------------:| 24 | | Data for Development (D4D) | 50,000 | 2 weeks | Ivory Coast | Available | | 25 | | SoBigData Tuscany CDRs | 860,000 | 1 month | Pisa, Lucca, Livorno, Firenze (Italy) | On-site Upon Project Approval | [https:bit.ly/CDRTuscany](https://bit.ly/CDRTuscany) | 26 | | SoBigData Rome CDRs | - | 10 months | Rome (Italy) | On-site Upon Project Approval | [bit.ly/CDRRome](https://bit.ly/CDRRome ) | 27 | | SoBigData Pisa CDRs | 230,000 | 1 month | Pisa (Italy) | On-site Upon Project Approval | [bit.ly/3gq9zUP](https://bit.ly/3gq9zUP) | 28 | | Telecom Big Data Challenge | - | 2 months | Milan and Trentino (Italy) | Available | [bit.ly/TBDC-2](https://bit.ly/TBDC-2) | 29 | | WorldPop Population Distribution | - | 2000-2020 | World | Available | [bit.ly/WorldPop-Data](https://bit.ly/WorldPop-Data) | 30 | | Changchun Temporal Distribution | 2 Million | 1 week | Changchun (China) | Available | [bit.ly/Changchun-Data](https://bit.ly/Changchun-Data) | 31 | | Songliao Basin Temporal Distr. | 3 Million | 1 week | Sangliao Basin Area (China) | Available | [bit.ly/Songliao](https://bit.ly/Songliao) | 32 | 33 | # Social Media 34 | 35 | | **Name** | **#Subjects** | **Timespan** | **Area** | **Access** | **Link** | 36 | |:-------------------------:|:-------------:|:------------:|:------------------------:|:-----------------:|:--------------------------------------------------:| 37 | | Gowalla | 190,000 | 20 months | California, Nevada (USA) | Available | [bit.ly/GowallaData](https://bit.ly/GowallaData) | 38 | | Brightkite | 58,000 | 30 months | - | Available | [bit.ly/Brightkite](https://bit.ly/Brightkite) | 39 | | Foursquare NYC, Singapore | 800,000 | 10 months | - | Available | [bit.ly/NYCSingapore](https://bit.ly/NYCSingapore) | 40 | | Yelp Open Dataset | 200,000 | - | - | Upon Registration | [bit.ly/YelpData](https://bit.ly/YelpData) | 41 | | YFCC100M | 100 Million | 10 years | - | Upon Request | [bit.ly/YFCC100M](https://bit.ly/YFCC100M) | 42 | 43 | # Others 44 | 45 | | **Name** | **#Subjects** | **Timespan** | **Area** | **Access** | **Link** | 46 | |:-------------------------:|:-------------:|:------------:|:---------------------:|:-----------------:|:------------------------------------------------------:| 47 | | Global Terrorism Database | - | From 1970 | World | Upon Registration | [bit.ly/GTDData](https://bit.ly/GTDData) | 48 | | Bike New York City | - | From 2013 | New York City (USA) | Available | [bit.ly/BikeNYCDAta](https://bit.ly/BikeNYCData) | 49 | | Bike Washington | - | From 2010 | Washington D.C. (USA) | Available | [bit.ly/BikeWashington](https://bit.ly/BikeWashington) | 50 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # [Deep Learning for Human Mobility: a Survey on Data and Models](https://arxiv.org/abs/2012.02825) 2 | 3 | 4 | 5 | This document aims to track the progress in the usage of Deep Learning (DL) applied to human mobility and give an overview of the state-of-the-art across the most common tasks and their corresponding datasets. In particular, we want to provide the users with a list of papers and they key characteristics (e.g., DL component(s) used in the model, metric(s) adopted for the evaluation and others) and, whenever it is open, a link to the dataset used in the paper. 6 | 7 | Moreover, we provide a list of datasources that can be used to model human mobility (e.g., Call Detail Records, GPS trajectories, Location Based Social Networks) and others that are not representing mobility but are strictly related to it and may be taken into account to finetune predictions (e.g., weather conditions, traffic data and others). 8 | 9 | This repository is based on the findings discussed in *Deep Learning for Human Mobility: a Survey on Data and Models* a paper by Massimiliano Luca, Gianni Barlacchi, Bruno Lepri and Luca Pappalardo. If you use the content of this repository or if you want to further investigate this topic using the survey, please cite our work as 10 | 11 | ```bibtex 12 | @misc{luca2020deep, 13 | title={Deep Learning for Human Mobility: a Survey on Data and Models}, 14 | author={Massimiliano Luca and Gianni Barlacchi and Bruno Lepri and Luca Pappalardo}, 15 | year={2020}, 16 | eprint={2012.02825}, 17 | archivePrefix={arXiv} 18 | } 19 | ``` 20 | 21 | __Table of Contents__ 22 | - [Datasets](https://github.com/scikit-mobility/DeepLearning4HumanMobility/tree/master/Datasets) 23 | - [Human Mobility](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md) 24 | - [Mobile Phone](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#mobile-phone) 25 | - [GPS Traces](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#gps) 26 | - [Social Media](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#social-media) 27 | - [Others](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/human_mobility.md#others) 28 | - [Auxiliary](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md) 29 | - [Spatial Aggregations](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#spatial-aggregations) 30 | - [Traffic](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#traffic) 31 | - [Environment](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#environment) 32 | - [Census and Administrative](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Datasets/auxiliary.md#census-and-administrative) 33 | - [Tasks](https://github.com/scikit-mobility/DeepLearning4HumanMobility/tree/master/Tasks) 34 | - [Next-Location Prediction](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Tasks/next-location.md) 35 | - [Crowd Flow Prediction](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Tasks/crowd.md) 36 | - [Trajectory Generation](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Tasks/generative.md) 37 | 38 | If you want to contribute to this repository, please open an issue and include the information according to the following guideline: 39 | 40 | __For a paper__ 41 | - The BibTex of the paper 42 | - The name of the model proposed 43 | - The DL components adopted in the model (e.g., LSTM, CNN, Convolutional Layers, Attention,...) 44 | - The evaluation metrics adopted 45 | - A link to an open datasets - if any 46 | - A link to a code repository - if any 47 | 48 | __For a dataset__ 49 | - The BibTex of the publication or the link related to the datasets. Please, do not use BibTex publications that just use the dataset. 50 | - Dataset type: auxiliary / core 51 | - The spatial and temporal coverage (e.g., Italy from Jan. 2020 to Apr. 2020) 52 | - The number of subjects covered (e.g., 500 taxis, 100k posts) 53 | - The accessibility conditions: available / on-site upon project approval / upon registration 54 | - A link to the dataset 55 | -------------------------------------------------------------------------------- /Tasks/README.md: -------------------------------------------------------------------------------- 1 | # Tasks - Deep Learning 4 Human Mobility 2 | #### [Back to Table Of Contents](https://github.com/scikit-mobility/DeepLearning4HumanMobility) 3 | --- 4 | 5 | In this part of the repository, you can access the state-of-art papers associated to each human mobility task under analysis. 6 | In particular, you can access 7 | 8 | - [Next-Location Prediction](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Tasks/next-location.md) 9 | - [Crowd Flow Prediction](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Tasks/crowd.md) 10 | - [Generative Models](https://github.com/scikit-mobility/DeepLearning4HumanMobility/blob/master/Tasks/generative.md) 11 | -------------------------------------------------------------------------------- /Tasks/crowd.md: -------------------------------------------------------------------------------- 1 | # Crowd Flow Prediction 2 | 3 | | **Paper** | **Model Name** | **Year** | **Model** | **Evaluation** | **Dataset** | **Code** | 4 | |:-----------------------------------------------------------------------------------------------------------------------------------------:|:--------------:|:--------:|:--------------------------------------:|:---------------:|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:--------------------------------------------:| 5 | | *Ren et al.,* A hybrid integrated deep learning model for the prediction of citywide spatio-temporal flow volumes | HIDLST | 2020 | LSTM, ST-ResNet | RMSE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | - | 6 | | *Tian et al.,* Deep spatial-temporal networks for crowd flows prediction by dilated convolutions and region-shifting attention mechanism. | LDRSN | 2020 | Convolutions, Attention Mechanism | RMSE, MAPE, MAE | [bit.ly/BikeNYCData](https://bit.ly/BikeNYCData) [bit.ly/TaxiNYC-2](https://bit.ly/TaxiNYC-2) | - | 7 | | *Yuan et al.,* Deep Multi-View Residual Attention Network for Crowd Flows Prediction | MV-RANet | 2020 | Residual Attention Network | RMSE, MAPE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | - | 8 | | *Liu et al.,* Dynamic Spatial-Temporal Representation Learning for Traffic Flow Prediction | ATFM | 2020 | ConvLSTM | RMSE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | [bit.ly/ATFM-2](https://bit.ly/ATFM-2) | 9 | | *Li et al.,* Densely Connected Convolutional Networks With Attention LSTM for Crowd Flows Prediction | ST-DCCNAL | 2019 | Densely Connected CNNs, Attention LSTM | RMSE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | [bit.ly/ST-DCCNAL](https://bit.ly/ST-DCCNAL) | 10 | | *Sun et al.,* Predicting citywide crowd flows in irregular regions using multi-view graph convolutional network | MVGCN | 2019 | Graph Convolutional Network | RMSE, MAE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) [bit.ly/BikeNYCData](https://bit.ly/BikeNYCData) [bit.ly/BikeWashington](https://bit.ly/BikeWashington) [bit.ly/TaxiNYC-2](https://bit.ly/TaxiNYC-2) | - | 11 | | *Lin et al.,* Deepstn+: Context-aware spatial-temporal neural network for crowd flow prediction in metropolis | DeepSTN+ | 2019 | Convolution, Residual Units | RMSE, MAE | [bit.ly/BikeNYCData](https://bit.ly/BikeNYCData) | [bit.ly/DeepSTN](https://bit.ly/DeepSTN) | 12 | | *Du et al.,* Deep irregular convolutional residual LSTM for urban traffic passenger flows prediction | DST-ICRL | 2019 | Convolutional Residual Units, LSTM | RMSE, MAE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | [bit.ly/DST-ICRL](https://bit.ly/DST-ICRL) | 13 | | *Ai et al.,* A deep learning approach on short-term spatiotemporal distribution forecasting of dockless bike-sharing system. | - | 2018 | Convolutional LSTM | RMSE, MAE | - | - | 14 | | *Yao et al.,* Revisiting Spatial-Temporal Similarity: A Deep Learning Framework for Traffic Prediction | STDN | 2018 | CNN, LSTM, Attention | RMSE, MAPE | [bit.ly/BikeNYCData](https://bit.ly/BikeNYCData) | [bit.ly/STDN-2](https://bit.ly/STDN-2) | 15 | | *Jin et al.,* Spatio-Temporal Recurrent Convolutional Networks for Citywide Short-Term Crowd Flows Prediction | STRCNs | 2018 | CNN, LSTM | RMSE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | - | 16 | | *Zonoozi et al.,* Periodic-CRN: A Convolutional Recurrent Model for Crowd Density Prediction with Recurring Periodic Patterns | PCRN | 2018 | Convolutional GRU. Convolutions | RMSE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | - | 17 | | *Want et al.,* Cross-city transfer learning for deep spatio-temporal prediction | RegionTrans | 2018 | Convolutional LSTM, CNN | RMSE | - | - | 18 | | *Zhang et al.,* Deep spatio-temporal residual networks for citywide crowd flows prediction | ST-ResNet | 2017 | Convolutions, Residual Units | RMSE | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) | [bit.ly/ST-ResNet](https://bit.ly/ST-ResNet) |s 19 | -------------------------------------------------------------------------------- /Tasks/generative.md: -------------------------------------------------------------------------------- 1 | | **Paper** | **Model Name** | **Year** | **Model** | **Evaluation** | **Dataset** | **Code** | 2 | |:-------------------------------------------------------------------------------------------------------------------------:|:--------------:|:--------:|:------------------------:|:----------------------------------------------:|:------------------------:|:----------------------------------------:| 3 | | *Berke et al.* Generating synthetic mobility data for a realistic population with {RNNs} to improve utility and privacy | - | 2022 | GAN, LSTM, RNN | KL on Distance, Locations per User, Aggregate Time per Location | - | [https://github.com/aberke/lbs-data/tree/master/trajectory_synthesis](github.com/aberke/lbs-data/tree/master/trajectory_synthesis) | 4 | | *Zhan et al.* Privacy-Aware Human Mobility Prediction via Adversarial Networks | LSTM-PAE | 2022 | AE, LSTM | Accuracy, Information Loss in Recostruction Process, User-re Identification Inaccuracy | [bit.ly/Geolife](https://bit.ly/Geolife) [bit.ly/MDC-2](https://bit.ly/MDC-2) [bit.ly/Foursquare-Data](https://bit.ly/Foursquare-Data) | - | 5 | | *Feng et al.* Learning to Simulate Human Mobility | MoveSim | 2020 | GAN, self-attention, CNN | Distance, rg, p(r,d), DailyLoc, G-rank, I-rank | [bit.ly/Geolife](https://bit.ly/Geolife) | [bit.ly/MoveSim](https://bit.ly/MoveSim) | 6 | | *Huang et al.* A Variational Autoencoder Based Generative Model of Urban Human Mobility | SVAE | 2019 | VAE, LSTM | MDE | - | - | 7 | | *Ouyang et al.* A Non-Parametric Generative Model for Human Trajectories | Ouyang GAN | 2018 | WGAN, CNN | | [bit.ly/MDC-2](https://bit.ly/MDC-2) | - | 8 | | *Kulkarni et al.* Generative models for simulating mobility trajectories | - | 2018 | RNN, GAN, copula | Statistical similarity, privacy test | [bit.ly/MDC-2](https://bit.ly/MDC-2) | - | 9 | | *Yin et al.* GANs based density distribution privacy-preservation on mobility data | - | 2018 | GAN, FC | Reconstruction error, Utility loss | [bit.ly/TaxiSF](https://bit.ly/TaxiSF) | - | 10 | | *Liu et al.* trajGANs: Using generative adversarial networks for geo-privacy protection of trajectory data (Vision paper) | trajGAN | 2018 | GANs | 11 | 12 | -------------------------------------------------------------------------------- /Tasks/next-location.md: -------------------------------------------------------------------------------- 1 | | **Paper** | **Model Name** | **Year** | **Model** | **Evaluation** | **Dataset** | **Code** | 2 | |:--------------------------------------------------------------------------------------------------------------:|:--------------:|:--------:|:----------------:|:----------------------:|:---------------------------------------------------------------------------------:|:------------------------------------------------:| 3 | | *Chen et al.* Context-aware Deep Model for Joint Mobility and Time Prediction | DeepJMT | 2020 | GRU, FC, Encoder | ACC@k | [bit.ly/Foursquare-Data](https://bit.ly/Foursquare-Data) | | 4 | | *Yang et al.* Location Prediction over Sparse User Mobility Traces Using RNNs: Flashback in Hidden States | Flashback | 2020 | Attention, RNN | ACC@k | [bit.ly/GowallaData](https://bit.ly/GowallaData) | [bit.ly/Flashback-1](https://bit.ly/Flashback-1) | 5 | | *Ebel et al.* Destination Prediction Based on Partial Trajectory Data | - | 2020 | RNN | Distance | [bit.ly/TaxiPorto](https://bit.ly/TaxiPorto), [bit.ly/TaxiSF](https://bit.ly/TaxiSF) | | 6 | | *Rossi et al.* Modelling Taxi Drivers’ Behaviour for the Next Destination Prediction | - | 2019 | Attention, LSTM | Distance | [bit.ly/TaxiPorto](https://bit.ly/TaxiPorto), [bit.ly/TaxiSF](https://bit.ly/TaxiSF), [bit.ly/TaxiNYC-2](https://bit.ly/TaxiNYC-2) | | 7 | | *Gao et al.* Predicting human mobility via variational attention | VANext | 2019 | CNN, GRU | ACC@k | [bit.ly/GowallaData](https://bit.ly/GowallaData) | | 8 | | *Kong et al.* HST-LSTM: A Hierarchical Spatial-Temporal Long-Short Term Memory Network for Location Prediction | HST-LSTM | 2018 | LSTM | ACC | - | [bit.ly/HST-LSTM](https://bit.ly/HST-LSTM) | 9 | | *Lv et al.* T-CONV: A convolutional neural network for multi-scale taxi trajectory prediction | T-VONC | 2018 | CNN | Distance | [bit.ly/TaxiPorto](https://bit.ly/TaxiPorto) | [bit.ly/T-CONV](https://bit.ly/T-CONV) | 10 | | *Feng et al.* Deepmove: Predicting human mobility with attentional recurrent networks | DeepMove | 2018 | Attention, RNN | ACC | [bit.ly/DeepMove](https://bit.ly/DeepMove) | [bit.ly/DeepMove](https://bit.ly/DeepMove) | 11 | | *Yao et al.* Revisiting Spatial-Temporal Similarity: A Deep Learning Framework for Traffic Prediction | SERM | 2017 | LSTM | ACC@k | - | [bit.ly/SERM-Repo](https://bit.ly/SERM-Repo) | 12 | | *Liu et al.* Predicting the next location: A recurrent model with spatial and temporal contexts | ST-RNN | 2016 | RNN | Rec@k, F1@k, MAPE, AUC | [bit.ly/GowallaData](https://bit.ly/GowallaData), [bit.ly/GTD](https://bit.ly/GTD) | [bit.ly/STRNN](https://bit.ly/STRNN) | 13 | | *De Brébisson et al.* Artificial neural networks applied to taxi destination prediction | - | 2015 | FC | Distance | [bit.ly/TaxiPorto](https://bit.ly/TaxiPorto) | [bit.ly/next-loc-1](https://bit.ly/next-loc-1) | 14 | --------------------------------------------------------------------------------