A Novel Approach to Mining Travel Sequences Using Collections of Geotagged Photos

Kisilevich, Slava; Keim, Daniel A.; Rokach, Lior

doi:10.1007/978-3-642-12326-9_9

Cited by 35 publications

(29 citation statements)

References 13 publications

(10 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Therefore, number of trajectories can be an index of popularity. In addition, stops are the most important parts of each trajectory (Spaccapietra et al ; Kisilevich ). These are places in which a driver shows a different behavior, such as slowing down or stopping for a while (Ashbrook and Starner ; Monajemi ).…”

Section: Enriching Road Segments Based On Multi‐domain Uggcmentioning

confidence: 99%

“…These data have a user-centric nature, and each data item indicates people's experiences, perceptions and emotions (Sester et al 2014). Most UGC data are located along road segments because they are produced during travels and trips (Popescu, Grefenstette, and Mo€ ellic 2009;Kisilevich, Keim, and Rokach 2010;Majid et al 2013). For example, people might take beautiful scenery pictures (Okuyama and Yanai 2013) or save their points of interest (POIs) spatial coordinates by using Global Positioning System (GPS) devices.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi‐Domain User‐Generated Content Based Model to Enrich Road Network Data for Multi‐Criteria Route Planning

Khoshamooz

Taleai

2017

Geographical Analysis

View full text Add to dashboard Cite

By utilizing today's web‐based technologies, people can act as sensors and share their perceptions, emotions and observations in a variety of data forms, such as images, videos, texts, Global Positioning System (GPS) trajectories and maps. These forms are collectively called user‐generated content (UGC). These data are in different domains and have a multi‐modality nature. Although recent efforts have probed the acquisition of local knowledge by using single‐domain UGC data in specific applications, such efforts have not thus far presented a model considering multi‐domain UGC specifically to enrich road network data. This article aims at presenting such a model wherein, with the help of each data domain of UGC, one aspect of people knowledge about the road segment is obtained. These different aspects of knowledge are integrated using a Skyline operator to support multi‐criteria route finding. We name this model ERSBU (enriching road segments based on UGC). In ERSBU, road segments are basic spatial units, and their subjective properties have been extracted by using available UGC. The scenic score for each road segment was computed by using geo‐tagged Panoramio photos. The accessibility level of each road segment to different facilities was calculated based on data captured from Wikimapia and OpenStreetMap. Moreover, for measuring the movement popularity of each road segment, Wikiloc and Everytrail GPS trajectories were utilized. For the implementation of the ERSBU model, Tehran region 6 was considered the case study area. The Evaluation of the results proved that road segments that achieved a high score based on knowledge extracted from UGC also mostly gained top scores by analyzing traditional maps. ERSBU allows users to accomplish more‐qualitative path finding by considering the multi‐view characteristics of road segments.

show abstract

Section: Enriching Road Segments Based On Multi‐domain Uggcmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Multi‐Domain User‐Generated Content Based Model to Enrich Road Network Data for Multi‐Criteria Route Planning

Khoshamooz

Taleai

2017

Geographical Analysis

View full text Add to dashboard Cite

show abstract

“…The significance of these data has been already addressed in the following papers as landmark identification [13,4], tag representative [1] and representative images [18]. However, since most of the research concentrates on application-based perspectives, systematic approaches for analysis of spatial or spatio-temporal aspects of geo-tagged photos only begin to emerge [8,12,15,14].…”

Section: Introductionmentioning

confidence: 99%

P-Dbscan

Kisilevich

Mansmann

Keim

2010

Proceedings of the 1st International Conference and Exhibition on Computing for Geospatial Research &Amp; Application

Self Cite

140

View full text Add to dashboard Cite

The rapid spread of location-based devices and cheap storage mechanisms, as well as fast development of Internet technology, allowed collection and distribution of huge amounts of user-generated data, such as people's movement or geotagged photos. These types of data produce new challenges for research in different application domains. In many cases, new algorithms should be devised to better portray the phenomena under investigation. In this paper, we present P-DBSCAN, a new density-based clustering algorithm based on DBSCAN for analysis of places and events using a collection of geo-tagged photos. We thereby introduce two new concepts: (1) density threshold, which is defined according to the number of people in the neighborhood, and (2) adaptive density, which is used for fast convergence towards high density regions. Our approach is demonstrated on the area of Washington, D.C.

show abstract

“…For example, the method can be used to analyze the trajectory of tourists in a new city they are visiting (Kisilevich et al, 2010). In such cases the trajectory is represented as a sequence of point-ofinterests which a tourist would explore.…”

Section: User Attributesmentioning

confidence: 99%

Attribute-Driven Hidden Markov Model Trees for Intention Prediction

Antwarg

Rokach

Shapira

2012

IEEE Trans. Syst., Man, Cybern. C

View full text Add to dashboard Cite

In this paper we introduce a novel approach for generating an intention prediction model of user interactions with systems. As part of this new approach, we include personal aspects such as user characteristics that can increase prediction accuracy. The model is automatically trained according to the user's fixed attributes (e.g., demographic data such as age and gender) and the user's sequences of actions in the system. The generated model has a tree structure. The building blocks of each node can be any probabilistic sequence model (such as hidden Markov models and conditional random fields)) and each node is split according to user attributes. Thus, we refer to this algorithm as an attribute-driven model tree. The new model was first tested on simulated data in which users with different attributes (such as age, gender) behave differently when trying to accomplish various tasks. We then validated the ability of the algorithm to discover the relevant attributes. We tested our algorithm on two real datasets: from a Web application and a mobile application dataset. The results were encouraging and indicate the capability of the proposed method for discovering the correct user intention model and increasing intention prediction accuracy compared to single HMM or CRF models. The motivation of this paper is to achieve a higher level of compatibility between user intentions and the system she is using. This motivation derives from the fact that in recent years applications have continuously been gaining functionality and emerging features are complicating user interactions with applications. Using intention prediction methods will enable a company to improve the services provided to users by adjusting the system services to their needs (e.g., the interface). Predicting user intentions can be also used for assisting users to perform or complete tasks, to alert them regarding the availability of features, and to increase their general knowledge. Effective assistance will allow users to acquire the new skills and knowledge they need to easily operate an unfamiliar device or function. Different users tend to select different sequences for accomplishing the same goal. Specifically, our hypothesis is that the user's attributes and context (such as age or the operating system) indicate which sequence the user will eventually perform. This hypothesis is based on studies that examined the interaction of users with systems (such as Web browsers) and found that user attributes affect the way of interaction. [For example, Hu et al., 2007, Weber and Castillo 2010; Thakor et al., 2004]. If a young male uses a system, it is possible to use the model to predict the goal that he intends to accomplish and provide him with a user interface (UI) more appropriate to his usage and intentions. Although methods exist for predicting user intentions, they do not take into account user attributes that can increase the accuracy of prediction. In this work we present a new method that we developed that examines how employing user attributes can...

show abstract

A Novel Approach to Mining Travel Sequences Using Collections of Geotagged Photos

Cited by 35 publications

References 13 publications

Multi‐Domain User‐Generated Content Based Model to Enrich Road Network Data for Multi‐Criteria Route Planning

Multi‐Domain User‐Generated Content Based Model to Enrich Road Network Data for Multi‐Criteria Route Planning

P-Dbscan

Attribute-Driven Hidden Markov Model Trees for Intention Prediction

Contact Info

Product

Resources

About