Qunying Huang, Zhenglong Li, Jing Li, Charles Chang
Trajectory pattern mining has been performed over many datasets, including animal movement, GPS trajectories, and human travel history. This paper aims to explore and mine individual frequently visited regions and trajectory patterns using online footprints captured through a social media site (i.e., Twitter). Regions of frequent visits representing daily activity areas at which an individual appears are derived using the DBSCAN clustering algorithm. A trajectory pattern mining algorithm is then applied to discover ordered sequences of these spatial regions that the individual visits frequently. To illustrate and test the effectiveness of the proposed methods, we analyze the activity patterns of a selected Twitter user using the geo-tagged tweets posted by the user for an extended period. The preliminary assessment indicates that our approach can be applied to mine individual frequent trajectory patterns from online footprints that are of relatively low and irregular spatial and temporal resolutions.
{"title":"Mining frequent trajectory patterns from online footprints","authors":"Qunying Huang, Zhenglong Li, Jing Li, Charles Chang","doi":"10.1145/3003421.3003431","DOIUrl":"https://doi.org/10.1145/3003421.3003431","url":null,"abstract":"Trajectory pattern mining has been performed over many datasets, including animal movement, GPS trajectories, and human travel history. This paper aims to explore and mine individual frequently visited regions and trajectory patterns using online footprints captured through a social media site (i.e., Twitter). Regions of frequent visits representing daily activity areas at which an individual appears are derived using the DBSCAN clustering algorithm. A trajectory pattern mining algorithm is then applied to discover ordered sequences of these spatial regions that the individual visits frequently. To illustrate and test the effectiveness of the proposed methods, we analyze the activity patterns of a selected Twitter user using the geo-tagged tweets posted by the user for an extended period. The preliminary assessment indicates that our approach can be applied to mine individual frequent trajectory patterns from online footprints that are of relatively low and irregular spatial and temporal resolutions.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132170836","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper we systematically explore aggregate operations over moving regions. We propose new aggregate operations and organize the operations according to their spatial, temporal, or spatiotemporal focus.
{"title":"Categorizing spatiotemporal aggregates for moving regions","authors":"Mark McKenney, Khalil Khobrani, Pr Rangaraju","doi":"10.1145/3003421.3003430","DOIUrl":"https://doi.org/10.1145/3003421.3003430","url":null,"abstract":"In this paper we systematically explore aggregate operations over moving regions. We propose new aggregate operations and organize the operations according to their spatial, temporal, or spatiotemporal focus.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130277481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
James N. Hughes, Matthew D. Zimmerman, Christopher N. Eichelberger, Anthony D. Fox
With the rise of location-aware IoT devices, there is an increased desire to process data streams in a real-time manner. Responding to such streams may require processing data from multiple streams to inform decisions. There are many uses cases for putting the location data from the sensors or an analytic derivative on a map for a live view of sensors or other assets. Here we describe an architecture which relies solely on free and open-source components to provide streaming spatio-temporal event processing, analysis, and near-real-time visualization.
{"title":"A survey of techniques and open-source tools for processing streams of spatio-temporal events","authors":"James N. Hughes, Matthew D. Zimmerman, Christopher N. Eichelberger, Anthony D. Fox","doi":"10.1145/3003421.3003432","DOIUrl":"https://doi.org/10.1145/3003421.3003432","url":null,"abstract":"With the rise of location-aware IoT devices, there is an increased desire to process data streams in a real-time manner. Responding to such streams may require processing data from multiple streams to inform decisions. There are many uses cases for putting the location data from the sensors or an analytic derivative on a map for a live view of sensors or other assets. Here we describe an architecture which relies solely on free and open-source components to provide streaming spatio-temporal event processing, analysis, and near-real-time visualization.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123913407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Spatio-temporal clustering aims to discover interesting regions in spatio-temporal data. In this paper, we propose a novel, serial, density-contour based spatio-temporal clustering algorithm called ST-DCONTOUR which employs a model-based clustering methodology to obtain spatio-temporal clusters from location streams. Our approach subdivides the incoming data into batches and employs a serial approach that generates spatial clusters for each batch first; next, spatio-temporal clusters are formed by identifying continuing relationships between spatial clusters in consecutive batches. Our approach employs contouring algorithms to identify spatial clusters as closed contours of a region where density is above a given threshold, and relies on contour analysis techniques to identify continuing, disappearing, and newly appearing spatial clusters in consecutive batches. We evaluate our approach by conducting a case study involving NYC taxi trips data. The experimental results show that ST-DCONTOUR can discover interesting spatio-temporal patterns in taxi pickup location streams.
{"title":"ST-DCONTOUR: a serial, density-contour based spatio-temporal clustering approach to cluster location streams","authors":"Yongli Zhang, C. Eick","doi":"10.1145/3003421.3003429","DOIUrl":"https://doi.org/10.1145/3003421.3003429","url":null,"abstract":"Spatio-temporal clustering aims to discover interesting regions in spatio-temporal data. In this paper, we propose a novel, serial, density-contour based spatio-temporal clustering algorithm called ST-DCONTOUR which employs a model-based clustering methodology to obtain spatio-temporal clusters from location streams. Our approach subdivides the incoming data into batches and employs a serial approach that generates spatial clusters for each batch first; next, spatio-temporal clusters are formed by identifying continuing relationships between spatial clusters in consecutive batches. Our approach employs contouring algorithms to identify spatial clusters as closed contours of a region where density is above a given threshold, and relies on contour analysis techniques to identify continuing, disappearing, and newly appearing spatial clusters in consecutive batches. We evaluate our approach by conducting a case study involving NYC taxi trips data. The experimental results show that ST-DCONTOUR can discover interesting spatio-temporal patterns in taxi pickup location streams.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129503387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Thanks to recent prevalence of location sensors, collecting massive spatiotemporal datasets containing moving object trajectories has become possible, providing an exceptional opportunity to derive interesting insights about behavior of the moving objects such as people, animals and vehicles. In particular, mining patterns from interdependent co-movements of objects in a group/team (such as players of a sports team, ants of a colony in search of food, and cars in a congested downtown district) can lead to the discovery of interesting patterns (e.g., offense tactics and strategies of a sports team). Various trajectory mining, and in particular frequent episode mining (FEM), approaches have been proposed to discover such patterns from trajectory datasets. However, the existing FEM approaches neither are applicable to multivariate spatial (MVS) event sequences nor consider and leverage all spatial features of the input data. In this paper, we first introduce a Spatial Apriori property which extends the well-known Apriori property to consider the spatial properties of the input data. We present a data preprocessing technique that leverages the aforementioned Spatial Apriori to reduce the search space of our problem by filtering out irrelevant events from a given MVS event sequence. Second, we present the MVS-FEM framework which efficiently discovers co-movements patterns from MVS datasets. The efficiency of our proposed solutions is evaluated using a real dataset.
{"title":"Mining frequent episodes from multivariate spatiotemporal event sequences","authors":"Shahab Helmi, F. Kashani","doi":"10.1145/3003421.3003428","DOIUrl":"https://doi.org/10.1145/3003421.3003428","url":null,"abstract":"Thanks to recent prevalence of location sensors, collecting massive spatiotemporal datasets containing moving object trajectories has become possible, providing an exceptional opportunity to derive interesting insights about behavior of the moving objects such as people, animals and vehicles. In particular, mining patterns from interdependent co-movements of objects in a group/team (such as players of a sports team, ants of a colony in search of food, and cars in a congested downtown district) can lead to the discovery of interesting patterns (e.g., offense tactics and strategies of a sports team). Various trajectory mining, and in particular frequent episode mining (FEM), approaches have been proposed to discover such patterns from trajectory datasets. However, the existing FEM approaches neither are applicable to multivariate spatial (MVS) event sequences nor consider and leverage all spatial features of the input data. In this paper, we first introduce a Spatial Apriori property which extends the well-known Apriori property to consider the spatial properties of the input data. We present a data preprocessing technique that leverages the aforementioned Spatial Apriori to reduce the search space of our problem by filtering out irrelevant events from a given MVS event sequence. Second, we present the MVS-FEM framework which efficiently discovers co-movements patterns from MVS datasets. The efficiency of our proposed solutions is evaluated using a real dataset.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127037704","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Accurate map matching has been a fundamental but challenging problem that has drawn great research attention in recent years. It aims to reduce the uncertainty in a trajectory by matching the GPS points to the road network on a digital map. Most existing work has focused on estimating the likelihood of a candidate path based on the GPS observations, while neglecting to model the probability of a route choice from the perspective of drivers. Here we propose a novel feature-based map matching algorithm that estimates the cost of a candidate path based on both GPS observations and human factors. To take human factors into consideration is very important especially when dealing with low sampling rate data where most of the movement details are lost. Additionally, we simultaneously analyze a subsequence of coherent GPS points by utilizing a new segment-based probabilistic map matching strategy, which is less susceptible to the noisiness of the positioning data. We have evaluated the proposed approach on a public large-scale GPS dataset, which consists of 100 trajectories distributed all over the world. The experimental results show that our method is robust to sparse data with large sampling intervals (e.g., 60 s ∼ 300 s) and challenging track features (e.g., u-turns and loops). Compared with two state-of-the-art map matching algorithms, our method substantially reduces the route mismatch error by 6.4% ∼ 32.3% and obtains the best map matching results in all the different combinations of sampling rates and challenging features.
准确的地图匹配是近年来备受关注的一个基础问题,也是一个具有挑战性的问题。它旨在通过将GPS点与数字地图上的道路网络相匹配来减少轨迹的不确定性。大多数现有的工作都集中在基于GPS观测估计候选路径的可能性上,而忽略了从驾驶员的角度建模路线选择的概率。本文提出了一种新的基于特征的地图匹配算法,该算法基于GPS观测值和人为因素来估计候选路径的成本。考虑人为因素是非常重要的,特别是在处理低采样率的数据时,大多数运动细节都丢失了。此外,我们还利用一种新的基于分段的概率地图匹配策略同时分析相干GPS点的子序列,该策略不易受定位数据噪声的影响。我们在一个公共的大规模GPS数据集上对所提出的方法进行了评估,该数据集由分布在世界各地的100条轨迹组成。实验结果表明,我们的方法对于大采样间隔(例如60 s ~ 300 s)和具有挑战性的轨迹特征(例如u形转弯和环路)的稀疏数据具有鲁棒性。与两种最先进的地图匹配算法相比,我们的方法将路径失配误差大幅降低了6.4% ~ 32.3%,并在所有不同采样率和挑战性特征的组合中获得了最佳的地图匹配结果。
{"title":"A general feature-based map matching framework with trajectory simplification","authors":"Yifang Yin, R. Shah, Roger Zimmermann","doi":"10.1145/3003421.3003426","DOIUrl":"https://doi.org/10.1145/3003421.3003426","url":null,"abstract":"Accurate map matching has been a fundamental but challenging problem that has drawn great research attention in recent years. It aims to reduce the uncertainty in a trajectory by matching the GPS points to the road network on a digital map. Most existing work has focused on estimating the likelihood of a candidate path based on the GPS observations, while neglecting to model the probability of a route choice from the perspective of drivers. Here we propose a novel feature-based map matching algorithm that estimates the cost of a candidate path based on both GPS observations and human factors. To take human factors into consideration is very important especially when dealing with low sampling rate data where most of the movement details are lost. Additionally, we simultaneously analyze a subsequence of coherent GPS points by utilizing a new segment-based probabilistic map matching strategy, which is less susceptible to the noisiness of the positioning data. We have evaluated the proposed approach on a public large-scale GPS dataset, which consists of 100 trajectories distributed all over the world. The experimental results show that our method is robust to sparse data with large sampling intervals (e.g., 60 s ∼ 300 s) and challenging track features (e.g., u-turns and loops). Compared with two state-of-the-art map matching algorithms, our method substantially reduces the route mismatch error by 6.4% ∼ 32.3% and obtains the best map matching results in all the different combinations of sampling rates and challenging features.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121773759","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we first categorize the various types of Network-constrained moving object queries. We then propose benchmarks that can be used to compare the performance of systems and indexing schemes that are proposed for handling these types of queries. Network-constrained moving objects are objects that move in a specific network, such as vehicles that are constrained to move in a road (traffic) network. Our query categories are based on the Network-constrained moving object model presented by [4, 6, 14]. We formally define comprehensive categories of typical queries, based on whether the conditions involve space (point versus region), time (point versus interval), and object id. The categories are based on the various combinations of these features. We describe the types of queries as Relational Calculus expressions, based on the query constraints. We focus on three main constraints: Spatial constraints, Temporal constraints, or/and moving object ID constraints. For each types of query, we identify the types of results, and give examples to clarify the query types. This work can define a benchmark for the performance of different types of systems and indexes that are designed to answer queries on Network-constrained moving objects data. Certain indexes/systems may work well for some query categories but perform poorly for other types of queries.
{"title":"Formalization of network-constrained moving object queries with application to benchmarking","authors":"M. Fouladgar, R. Elmasri","doi":"10.1145/3003421.3003427","DOIUrl":"https://doi.org/10.1145/3003421.3003427","url":null,"abstract":"In this paper, we first categorize the various types of Network-constrained moving object queries. We then propose benchmarks that can be used to compare the performance of systems and indexing schemes that are proposed for handling these types of queries. Network-constrained moving objects are objects that move in a specific network, such as vehicles that are constrained to move in a road (traffic) network. Our query categories are based on the Network-constrained moving object model presented by [4, 6, 14]. We formally define comprehensive categories of typical queries, based on whether the conditions involve space (point versus region), time (point versus interval), and object id. The categories are based on the various combinations of these features. We describe the types of queries as Relational Calculus expressions, based on the query constraints. We focus on three main constraints: Spatial constraints, Temporal constraints, or/and moving object ID constraints. For each types of query, we identify the types of results, and give examples to clarify the query types. This work can define a benchmark for the performance of different types of systems and indexes that are designed to answer queries on Network-constrained moving objects data. Certain indexes/systems may work well for some query categories but perform poorly for other types of queries.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115324124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The main objectives of moving objects queries are to search for objects that either lie in some specific areas (i.e., range queries) or are close to one specific location (i.e., kNN queries). Such queries have been previously studied considering either offline database processes using some index techniques or online approaches where incoming data are processed to answer those queries "on the fly". The research presented in this paper considers hybrid queries applied to historical data as well as streaming data. When considering the specific context of the maritime domain and moving objects at sea, a key issue is to make a difference between covered and non covered areas (i.e., regions from where AIS positioning signals are either received or not received). This leads us to introduce the concept of "Black Holes" query where the objective is to identify regions respectively covered and non covered, this providing useful insights for maritime authorities in charge of the regulation of maritime transportation.
{"title":"Continuous detection of black holes for moving objects at sea","authors":"Loïc Salmon, C. Ray, Christophe Claramunt","doi":"10.1145/3003421.3003423","DOIUrl":"https://doi.org/10.1145/3003421.3003423","url":null,"abstract":"The main objectives of moving objects queries are to search for objects that either lie in some specific areas (i.e., range queries) or are close to one specific location (i.e., kNN queries). Such queries have been previously studied considering either offline database processes using some index techniques or online approaches where incoming data are processed to answer those queries \"on the fly\". The research presented in this paper considers hybrid queries applied to historical data as well as streaming data. When considering the specific context of the maritime domain and moving objects at sea, a key issue is to make a difference between covered and non covered areas (i.e., regions from where AIS positioning signals are either received or not received). This leads us to introduce the concept of \"Black Holes\" query where the objective is to identify regions respectively covered and non covered, this providing useful insights for maritime authorities in charge of the regulation of maritime transportation.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131903251","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Samara Martins do Nascimento, J. Macêdo, Mirla Rafaela Rafael Braga Chucre, M. Casanova, Javam C. Machado
Time dependent networks in mobility scenario are key for many applications that need to cope with real world dynamics. However, the quality of a time dependent network relies on the accuracy of its temporal functions. To this aim, we propose a new method for computing temporal functions for a time dependent network using Trajectory Data Streams. This proposal extends the previous Piecewise linear model, which uses a smooth curve approach, called LOESS, that can estimate where the breakpoints values occurs in a Piecewise linear function. A challenge faced by the use of trajectory data streams is related with the time constraint to update time dependent network time functions. Our model computes the time dependent network and update the temporal function that needs to reflect recent data and discard old data. We described our solution and present experimental results, which show that our approach is efficient and effective comparing to their competitors.
{"title":"On computing temporal functions for time-dependent networks using trajectory data streams","authors":"Samara Martins do Nascimento, J. Macêdo, Mirla Rafaela Rafael Braga Chucre, M. Casanova, Javam C. Machado","doi":"10.1145/3003421.3003425","DOIUrl":"https://doi.org/10.1145/3003421.3003425","url":null,"abstract":"Time dependent networks in mobility scenario are key for many applications that need to cope with real world dynamics. However, the quality of a time dependent network relies on the accuracy of its temporal functions. To this aim, we propose a new method for computing temporal functions for a time dependent network using Trajectory Data Streams. This proposal extends the previous Piecewise linear model, which uses a smooth curve approach, called LOESS, that can estimate where the breakpoints values occurs in a Piecewise linear function. A challenge faced by the use of trajectory data streams is related with the time constraint to update time dependent network time functions. Our model computes the time dependent network and update the temporal function that needs to reflect recent data and discard old data. We described our solution and present experimental results, which show that our approach is efficient and effective comparing to their competitors.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133287892","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mobility prediction is becoming one of the key elements of location-based services. In the near future, it will also facilitate tasks such as resource management, logistics administration and urban planning. To predict human mobility, many techniques have been proposed. However, existing techniques are usually driven by large volumes of data to train user mobility models computed over a long duration and stored in a centralized server. This results in inherently long waiting times before the prediction model kicks in. Over this large training data, small time bounded user movements are shadowed, due to their marginality, thus impacting the granularity of predictions. Transferring highly sensitive location data to third party entities also exposes the user to several privacy risks. To address these issues, we propose MobiDict, a realtime mobility prediction system that is constantly adapting to the user mobility behaviour, by taking into account the movement periodicity and the evolution of frequently visited places. Compared to the existing training approaches, our system utilises less data to generate the evolving mobility models, which in turn lowers the computational complexity and enables implementation on handheld devices, thus preserving privacy. We test our system using mobility traces collected around Lake Geneva region from 168 users and demonstrate the performance of our approach by evaluating MobiDict with six different prediction techniques. We find a satisfactory prediction accuracy as compared to the baseline results obtained with 70% of the user dataset for majority of the users.
{"title":"MobiDict: a mobility prediction system leveraging realtime location data streams","authors":"Vaibhav Kulkarni, A. Moro, B. Garbinato","doi":"10.1145/3003421.3003424","DOIUrl":"https://doi.org/10.1145/3003421.3003424","url":null,"abstract":"Mobility prediction is becoming one of the key elements of location-based services. In the near future, it will also facilitate tasks such as resource management, logistics administration and urban planning. To predict human mobility, many techniques have been proposed. However, existing techniques are usually driven by large volumes of data to train user mobility models computed over a long duration and stored in a centralized server. This results in inherently long waiting times before the prediction model kicks in. Over this large training data, small time bounded user movements are shadowed, due to their marginality, thus impacting the granularity of predictions. Transferring highly sensitive location data to third party entities also exposes the user to several privacy risks. To address these issues, we propose MobiDict, a realtime mobility prediction system that is constantly adapting to the user mobility behaviour, by taking into account the movement periodicity and the evolution of frequently visited places. Compared to the existing training approaches, our system utilises less data to generate the evolving mobility models, which in turn lowers the computational complexity and enables implementation on handheld devices, thus preserving privacy. We test our system using mobility traces collected around Lake Geneva region from 168 users and demonstrate the performance of our approach by evaluating MobiDict with six different prediction techniques. We find a satisfactory prediction accuracy as compared to the baseline results obtained with 70% of the user dataset for majority of the users.","PeriodicalId":210363,"journal":{"name":"Proceedings of the 7th ACM SIGSPATIAL International Workshop on GeoStreaming","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2016-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128668738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}