Predicting Human Mobility via Variational Attention

Gao, Qiang; Zhang, Fan; Trajcevski, Goce; Zhang, Kunpeng; Zhong, Ting; Zhang, Fengli

doi:10.1145/3308558.3313610

Cited by 106 publications

(46 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We apply Negative Log-Likelihood (NLL) [10], the prediction accuracy [3,16] and Average Percentage Rank [14,15] to evaluate the performance of our method. To demonstrate the effectiveness of STSAN, we compared with the following location prediction methods: ST-RNN [11], MCARNN [10], DeepMove [3], and VANext [4]. To demonstrate the effectiveness of federated learning model AMF, We compared with FedAvg [13] in our proposed STSAN method for location prediction task.…”

Section: Experiments 31 Experimental Settingsmentioning

confidence: 99%

Predicting Human Mobility with Federated Learning

Wang

et al. 2020

Proceedings of the 28th International Conference on Advances in Geographic Information Systems

View full text Add to dashboard Cite

In recent years, location prediction has become an important task and has gained significant attention. Existing location prediction methods rely on centralized storage of user mobility data for model training, which may lead to privacy concerns and risks due to the privacy-sensitive nature of user behaviors. In this work, we propose a privacy-preserving method for mobility prediction model training based on federated learning, which can leverage the useful information in the behaviors of massive users to train accurate mobility prediction models and meanwhile remove the need to centralized storage of them. Firstly, we propose a novel network named STSAN (Spatial-Temporal Self-Attention Network) on each user device, which can integrate spatiotemporal information with the self-attention for location prediction and a new personalized federated learning model named AMF (Adaptive Model Fusion Federated Learning), which is a mixture of local and global model. Finally, the results are superior to various baselines on four realworld check-ins datasets, verifying the effectiveness of the method. CCS CONCEPTS • Security and privacy → Human and societal aspects of security and privacy; • Information systems → Location based services; • Human-centered computing → Ubiquitous and mobile computing design and evaluation methods.

show abstract

Section: Experiments 31 Experimental Settingsmentioning

confidence: 99%

Predicting Human Mobility with Federated Learning

Wang

et al. 2020

Proceedings of the 28th International Conference on Advances in Geographic Information Systems

View full text Add to dashboard Cite

show abstract

“…Human mobility prediction has been widely studied in the filed of urban computing [48]. Many trajectory-based deep learning models were proposed to predict each individual's movement [6,8,25]. [25] extends a regular RNN by utilizing time and distance specific transition matrices to propose an ST-RNN model for predicting the next location.…”

Section: Urban Computingmentioning

confidence: 99%

“…DeepMove [6], considered to be a state-of-the-art model for trajectory prediction, designed a historical attention module to capture periodicities and augment prediction accuracy. VANext [8] further enhanced DeepMove by proposing a novel variational attention mechanism. Besides the prediction at an individual level, modeling and predicting millions of individuals' mobility at a citywide level has also been studied.…”

Section: Urban Computingmentioning

confidence: 99%

“…With the application and popularization of location-acquisition technologies, huge amounts of trajectory data are being generated and collected from various sources such as mobile phones, sensors, and car navigation systems for analyzing human mobility. By utilizing these, many researchers focus on pure human movement such as: (1) human mobility prediction/modeling from an individual level to a citywide level [4,6,8,25,32,41,43]; (2) human mobility pattern mining based on spatial and temporal similarity [9,34,35,38,49]. Meanwhile, higher research level aims to understand trajectory, saying, why people move to somewhere?…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DualSIN

Chen

Jiang

Yang

et al. 2020

Proceedings of the 28th International Conference on Advances in Geographic Information Systems

View full text Add to dashboard Cite

Nowadays, GPS devices have increased explosively and produced huge amounts of trajectory data related to people's outgoing. Through those big location data, many researches aim to analyze human mobility for urban development, such as human movement prediction/modeling, POI (Point-Of-Interest) recommendation. However, trajectory data only contains timestamp and location information. The intention of human movement is not explicit so that it is hard to understand why people go to somewhere. The intention prior to the activity could be of great significance for analyzing and predicting human mobility, which has not been taken into consideration by the existing researches until the present. Thus, in this study, we propose a brand-new concept called human intentional mobility, aiming to employ intention information to predict people's outgoing. We carefully utilize user's search query to sense his intention as well as the intensity. For instance, if a user searches a certain POI for many times in a short period, it will represent a relatively high intention to go there. Then, to fully utilize this intention representation for predicting whether user will visit searched POI or not, we specially design Dual Sequential Interaction Network (DualSIN) as a novel and unique deep-learning model, which can effectively capture the sophisticated interactions among two kinds of sequential information (i.e., search sequence and mobility sequence) and typical categorical information (i.e., user attributes). Last, we evaluate our model on real-world dataset collected from Yahoo! Japan portal application, and demonstrate that it can achieve superior satisfactory performances to the-state-of-the-art models on multiple POI search queries. CCS CONCEPTS • Information systems → Recommender systems; Information systems applications; • Human-centered computing → * Equal contribution.

show abstract

“…Both target is achieved successfully and provides a better performance than Pearson correlation-based CF model with the similar eectiveness. VANext (Variation Attention based Next) was discussed in [5], which captures sequential patterns of user check-in behavior for POI prediction. The proposed model is superior to the widely used RNN in that it can capture individual movement patterns more eciently.…”

Section: Current Studies On Location Based Recommendationsmentioning

confidence: 99%

A Proposal of Latent Interest Analysis by Geo-tagged SNS for Advertisement Recommendation

Omura

Kawai

Nakajima

et al. 2020

Proceedings of the 28th International Conference on Advances in Geographic Information Systems

View full text Add to dashboard Cite

advertisement (ad) recommendation services for mobile users is rapidly increasing. The conventional ways of recommending ads are based on the analysis of users' explicit behavior such as search keywords and keyword matching based on browsing history. However, it might not be eective enough for latent buyers. We have been working on an analysis of the user's latent interest on web browsing history which categorized positive and negative behaviors. In this paper, we adapt the method of the linked pages to real world locations using geo-tagged tweets. By several evaluations, we discuss the possibility to recommend ads according to the user's current location.

show abstract

Predicting Human Mobility via Variational Attention

Cited by 106 publications

References 23 publications

Predicting Human Mobility with Federated Learning

Predicting Human Mobility with Federated Learning

DualSIN

A Proposal of Latent Interest Analysis by Geo-tagged SNS for Advertisement Recommendation

Contact Info

Product

Resources

About