Figure 1 - uploaded by Dan Zou
Content may be subject to copyright.
Source publication
Many mobility prediction models have emerged over the past decade to predict a user’s next location through the utilisation of user trajectories. However, the performance is constrained by the quantity of user trajectory data. This research introduces a new approach by combining knowledge of individual travel behaviour and collective preferences of...
Contexts in source publication
Context 1
... and travel distance preferences). The workflow of the proposed model is shown in Figure 1. The three main components of the model are user clustering, Markov-based location prediction based on an individual's historical trajectories, and collective location preference extraction for each cluster of users. ...
Context 2
... comparing with the LSTM model here does not make sense. As shown in Figure 10, the two models (i.e. MUCS and MUCA) based on both the user's historical trajectories and the collective location preferences performed more robustly than MM. ...
Context 3
... the 'family person' group ( Figure 10(a)), the prediction accuracy of all three models was high even in the early stage, and it reached more than 93%. The limited historical trajectory had little impact on the next location prediction in this group. ...
Context 4
... we used a boxplot to show the distributions of α, β, and γ . Figure 11 shows the distributions of α, β, and γ from 1:00 to 24:00 of the 'regular worker' group (see Appendix 1 for distributions for 'family person', 'active family person', 'afternoon worker', and 'part-time worker'). ...
Context 5
... shown in Figure 11, the median values of α, β, and γ show distinct variation patterns. These parameters vary among users in different clusters and vary with time, which helps us understand how these behavioural factors impact an individual's next location prediction at different times. ...
Context 6
... is clearly shown in Figure 12 that the decrease in prediction accuracy is apparent for 'family person', 'regular worker', and 'afternoon worker' when β is 0. Take 'regular worker' as an example. The decline was concentrated when users' activities and locations changed, such as 9:00 and 10:00 in the morning and 19:00 and 20:00. ...
Similar publications
Human mobility plays an important role in the dynamics of infectious disease spread. Evidence from the initial nationwide lockdowns for COVID− 19 indicates that restricting human mobility is an effective strategy to contain the spread. While a direct correlation was observed early on, it is not known how mobility impacted COVID− 19 infection growth...
Citations
... From the architectural or pipeline perspective, some methods incorporate inverse reinforcement learning [19] or imitation learning [39,31] to capture the underlying mechanisms of human travel. Likewise, many articles draw inspiration from recommendation systems, suggesting that individuals who have similar travel patter are likely to visit similar locations in the future [8,4], besides, there are also some papers using knowledge graph to improve the prediction accuracy [38]. There is an abundance of literature in this area, and interested readers can refer to recent review papers for further exploration [11,9]. ...
Predicting future bus trip chains for an existing user is of great significance for operators of public transit systems. Existing methods always treat this task as a time-series prediction problem, but the 1-dimensional time series structure cannot express the complex relationship between trips. To better capture the inherent patterns in bus travel behavior, this paper proposes a novel approach that synthesizes future bus trip chains based on those from similar days. Key similarity patterns are defined and tested using real-world data, and a similarity function is then developed to capture these patterns. Afterwards, a graph is constructed where each day is represented as a node and edge weight reflects the similarity between days. Besides, the trips on a given day can be regarded as labels for each node, transferring the bus trip chain prediction problem to a semi-supervised classification problem on a graph. To address this, we propose several methods and validate them on a real-world dataset of 10000 bus users, achieving state-of-the-art prediction results. Analyzing the parameters of similarity function reveals some interesting bus usage patterns, allowing us can to cluster bus users into three types: repeat-dominated, evolve-dominate and repeat-evolve balanced. In summary, our work demonstrates the effectiveness of similarity-based prediction for bus trip chains and provides a new perspective for analyzing individual bus travel patterns. The code for our prediction model is publicly available.
... Wang et al. (2020) proposed an algorithm based on multi-task learning, which overcame the trajectory sparsity by jointly modeling the similarity and learning the mobility patterns of different individuals. Li et al. (2022) combined individual travel behaviors with the collective preferences of users sharing similar daily activity patterns for next location prediction, with up to 16% improvement in accuracy. These findings all show the advantages of utilizing other users' trajectories in the prediction for sparse trajectory data. ...
The spatiotemporal mobility patterns and next location prediction of fake base stations (FBS) provide important technical support for the police to prevent spam messages from FBS. However, due to the difficulty in locating their real-time locations, our understanding of the mobility patterns and predictability of FBS is still limited. Based on the crowdsourced spam data, we extract the time and potential locations of FBS and propose a Tucker-MMC method that combines Tucker decomposition with a Mobility Markov Chain (MMC) model to investigate the mobility patterns and predictability of FBS sending spam messages. First, we utilize Tucker decomposition to reflect the spatial and temporal preferences during the movement of the corresponding FBS. Then the mobility regularity and the theoretical maximum predictability of the FBS trajectories with similar mobility preferences are analyzed by entropy and Fano's inequality. A Tucker-MMC is also established for the next location prediction. The results using the spam dataset in Beijing show that the accuracy of Tucker-MMC is more than double that of the MMC. The accuracy of the actual location prediction model is more likely to approach the theoretical maximum predictability when FBS send spam messages in a shorter time, shorter transfer distance, and smaller access range.
... Individuals are unable to conceive of their existence without a smartphone or personal computer [3]. Technological advancements have yielded extensive trajectory data for studies and applications related to individual mobility [4]. Furthermore, examining the movement of individuals is essential for addressing other significant societal issues, such as urbanization, segregation, and the proliferation of epidemics, among others [5,6]. ...
Accurate prediction of activity location is a crucial component in various mobility applications and is particularly vital for the creation of customized, environmentally friendly transport systems. Next-location prediction, which entails predicting a user's forthcoming place by analyzing their previous movement patterns, has substantial ramifications in diverse fields, such as urban planning, geo-marketing, disease transmission, wireless network performance, recommender systems, and numerous other sectors. Recently, researchers have proposed a variety of predictors, including cutting-edge ones that utilize advanced deep learning methods, to tackle this problem. This study introduces robust models for predicting a user's future location based on their previous location. It proposes a Recurrent Neural Networks (RNNs) prediction scheme and a Gated Recurrent Unit (GRU), which are well-suited for learning from sequential data. Additionally, the clustering technique Density-Based Clustering (DBSCAN) is implemented to extract the stay points. Furthermore, the suggested method is more accurate at predicting the future than the current method, showing improvements in loss mean square error of up to 0.0005 in the RNN model and 0.01 in the GUR model. So, the models that were used led to a decrease in loss MSE, which was shown in the real-world dataset (Geolife) in this paper. The results are also consistent with other similar works that look at the same issue, showing how well the models can predict mobility.
... Human activities, including mandatory activities (e.g., work and school), maintenance activities (e.g., shopping, banking, doctor visits,) and leisure activities (e.g., visiting friends, sporting activities) (Lu and Pas, 1999;Sari Aslam et al., 2020), are of great importance for understanding travel behavior and urban dynamics (Calabrese et al., 2013;Rasouli and Timmermans, 2014;Li et al., 2022). Nowadays, people live more diverse lifestyles and spend a larger percentage of their time on non-mandatory (maintenance and leisure) activities than ever (Jim and Chen, 2009;Zhong et al., 2014). ...
Non-mandatory activities (e.g., shopping and leisure) are irregular in space and time, resulting in complex interactions between individuals and urban spaces. Understanding the associated factors of non-mandatory activities is vital for effective urban transport planning and management. This study uses travel survey data from Guangzhou, China, and a random forest (RF) model to investigate non-linear relationships between non-mandatory activities and their associated factors from the perspectives of time, location, built environment, activity dependency, and individual socioeconomic status, on both weekdays and weekends. The contribution of each factor to different non-mandatory activities is examined by a post hoc interpretable method, Shapley Additive exPlanations (SHAP). The results show that activity start time and activity dependency factors have a more significant impact on non-mandatory activities on weekdays, while duration has a greater influence on weekends. Built environment factors like wholesale and retail points of interest (POIs) play a significant role in shopping activities on both weekdays and weekends, while tourism POIs have a greater impact on leisure activities on weekends. Additionally, our analysis reveals the nonlinear dependencies and threshold effects of the top three factors for each category of non-mandatory activities and highlights their disparities between weekdays and weekends.
... For long-distance trajectory prediction, Wang et al. [28] proposed a multiuser multistep trajectory prediction method which incorporated long short-term memory (LSTM) and sequence-to-sequence (Seq2Seq) learning. For location prediction, Li et al. [29] proposed a prediction framework which integrated individual travel behavior and collective preferences for next location prediction. Different from the traditional trajectory prediction problem, the trajectory prediction of task offloading belongs to the node prediction problem. ...
Real-time computation tasks in vehicular edge computing (VEC) provide convenience for vehicle users. However, the efficiency of task offloading seriously affects the quality of service (QoS). The predictive-mode task offloading is limited by computation resources, storage resources and the timeliness of vehicle trajectory data. Meanwhile, machine learning is difficult to deploy on edge servers. In this paper, we propose a vehicle trajectory prediction method based on the vehicle frequent pattern for task offloading in VEC. First, in the initialization stage, a T-pattern prediction tree (TPPT) is constructed based on the historical vehicle trajectory data. Secondly, when predicting the vehicle trajectory, the vehicle frequent itemset with the largest vehicle trajectory support is found in the vehicle frequent itemset of the TPPT. Finally, in the update stage, the TPPT is updated in real time with the predicted vehicle trajectory results. Meanwhile, based on the proposed prediction method, the strategies of task offloading and optimization algorithm are designed to minimize energy consumption with time constraints. The experiments are carried out on real-vehicle datasets and the Capital Bikeshare datasets. The results show that compared with the baseline T-pattern method, the accuracy of the prediction method is improved by more than 10% and the prediction efficiency is improved by more than 6.5 times. The vehicle trajectory prediction method based on the vehicle frequent pattern has high accuracy and prediction efficiency, which can solve the problem of vehicle trajectory prediction for task offloading.
... As previously discussed, the number of clusters must be pre-specified. To do that, the Elbow method is used (Li, Zou, and Xu 2022). Figure 4 illustrates that the optimal number of clusters is determined to be 5. ...
The main objective of this study is to put forward a typology to describe better associations between road segments and driving patterns as reflected by driving speed. As the first step, a regression model is developed to examine the association between road segments and driving speed. Then, various unsupervised machine learning techniques, including k-means, AHC, and k-proto, are used to develop typologies of road segments. Speed data from a fleet of taxis operating in Montreal, Quebec, are used to validate the discrimination power of the various typologies. Results demonstrate that combining k-means and Gower distance produces the most accurate road typology. Various statistical tests, including ANOVA, Leven, and post hoc analyses, confirmed that the speed values of the various road types are significantly different. Finally, R² of regression models developed for various road types demonstrated that the generated road types better elucidate the variability of driving speed.
... Then, the ping-pong effects in the cellphone positioning data were eliminated using the point-clustering method proposed by Xu et al. (2020). Finally, the nearest neighbor interpolation method (Hoteit et al., 2014;Li et al., 2019Li et al., , 2021 was utilized to fill in the missing locations of the cellphone user trajectories at hourly intervals. Then, a missing record could be interpolated by the value of its nearest sampling position in time. ...
PM2.5 pollution imposes substantial health risks on urban residents. Previous studies mainly focused on assessing peoples' exposures at static locations, such as homes or workplaces. There has been a scarcity of research that quantifies the dynamic PM2.5 exposures of people when they travel in cities. To address this gap, we use cellphone positioning data and PM2.5 concentration data collected from smart sensors along roads in Guangzhou, China, to assess personal travel exposure to on-road PM2.5. First, we extract the trips of cellphone users from their trajectories and use the shortest path algorithm to calculate their travel routes on the road network. Second, the travel exposure of each user is estimated by associating their movement patterns with PM2.5 concentrations on roads. The result shows that most users’ average travel exposures per hour fall within the range of 20 ug/m³ to 75 ug/m³. Travel exposure varies across users, and 54.0% of users experience low travel exposure throughout the day, 25.5% of users experience high travel exposure in the evening, and 20.5% of users experience high travel exposure in the afternoon. Furthermore, the impacts of on-road PM2.5 on urban populations are uneven across roads. More attention should be given to roads with high PM2.5 concentrations and traffic flows in each period, such as Huan Shi Middle Road in the morning, Inner Ring Road in the afternoon, and Xinjiao Middle Road in the evening. The findings in this study can contribute to a more in-depth understanding of the relationship between air pollution and the travel activities of urban populations.
This paper examines the landscape of contactless transport payment systems in developing countries, focusing on data ownership, regulatory challenges, advantages, disadvantages, and implications. It explores the roles of city authorities, contractors, and consumers in managing and utilising data generated through these systems, emphasising the need for robust cybersecurity measures and comprehensive regulatory frameworks. Despite offering enhanced convenience and operational efficiencies, these systems present challenges such as cybersecurity risks, adoption barriers, and ethical concerns over data commercialisation. Plans include integrating these systems with international travel apps, developing user-friendly applications, and enhancing security protocols. The paper highlights the theoretical contributions and opportunities for future research, urging further exploration into regulatory effectiveness, consumer behaviour, technological advancements, ethical dilemmas, regional comparisons, and long-term impacts on urban mobility.
Accurately predicting human mobility is crucial for various applications, e.g., transportation services, epidemic control, and advertisement recommendation. Although numerous sequential modeling based methods (e.g., recurrent neural networks) have been proposed for human mobility prediction, accurately modeling individuals’ high-order travel preferences and the influence of social neighbors on their travel decisions remains challenging. In this paper, we construct a novel multi-context aware model for next location prediction, which aggregates multi-dimensional contextual features, including individual preferences, social relations, and activity-location associations. First, we define activity prediction as an auxiliary task and propose an activity-location association pruning method to mitigate the impact of data sparsity on model prediction. Second, we present a novel motif-preserving individual travel preference learning method that leverages a motif-induced hypergraph convolutional network to capture high-order travel preference features explicitly. Third, we identify virtual social neighbors with similar preferences based on individual travel preference learning results, and design a new social gated fusion structure to model the influence of social neighbors on individual travel choices. Finally, experimental results on two real-world travel datasets demonstrate the superiority of the proposed model over baseline models. Our proposed universal method can be seamlessly integrated with other sequential prediction models to improve the accuracy and stability of human mobility prediction.
