Towards a Semantic Approach for the Design of Social Network

Users’ Geographical Trajectories

Hadhami Ounissi

, Marwa Manaa

and Jalel Akaichi

Université de Tunis, ISG, BESTMOD, 41 Rue de La Liberté Cité Bouchoucha, 2000 Le Bardo, Tunis, Tunisia

College of Computer Science, King Khaled University, Abha, Saudi Arabia

Keywords: Human Mobility, Ontology, Semantic Modeling, Social Networks, Spatial Data, Trajectory Data.

Abstract: The volume of data keeps growing rapidly, especially with the arrival and the frequent access to social

networks. The spread of these networks provides users the opportunity to share their social, geographical and

temporal information through geo-localized tweets and check-ins. The challenge is to exploit these data leads

to a decision in favour of different situations encountered by these users. Thus, if we successfully analyze

their trends according to the models of users’ movements, we can then draw conclusions about the evolution

of their instantaneous behavior and accomplished activities. But, the problem is that the use of such data

decrees the provision of a representative formalism that combines spatial data and user information. In this

paper, we propose an approach for a semantic modeling of social network users’ trajectories. To do so,

ontology seems to be a promising solution that allows us to annotate raw trajectories with semantic

information to give birth to semantic trajectories. Such semantic trajectories are then analyzed in order to

detect user behavior in a dynamic way.

1 INTRODUCTION

These years were marked by the advent of web 2.0

and web 3.0. Such latest technologies have allowed

the influx of social networking. Hence, Facebook,

Twitter, LinkedIn, Google+, and many others are

considered as new darlings of Web 2.0 on the web.

Social networks are booming because they attract

more and more users and researchers. Today, they

allow users to share their personal information, their

desires and their photos, to communicate and record

movements tracks through GPS technology

integrated in mobile phones (smartphones, PDAs,

IPhone, etc.). This vigorous prosperity gave birth to

Twitter Places in June 2010 and Facebook Places end

of August 2010, where each includes its own

geolocation application. Moreover, this new

application collects the trajectory data generated by

users which are defined as traces of people in motion.

Therefore, data that convey these networks are huge

amounts concealing important conclusions. In fact,

this research area has persuaded a lot of researchers

to understand the user’s behavior by referring to their

movements.

Various problems are derived from this context.

(i) The first problem is to know how to benefit from

the great mass of geographic information. This

information is increasingly accessible and available

through the movement of persons registered at each

moment. Our interest is to get benefits not only from

this information but also to exploit other existing

information on these networks such as the metadata

i.e., name and type of the place, time, comments on a

subject, photos, mood and friendships. (ii) The

second problem is the need to identify, analyze and

understand the behavior of Internet users through

social networks by exploiting their trajectories.

Treatment of semantics related to trajectories will be

essential in order to achieve semantic behaviors.

Therefore, taking trajectories in their raw states is an

unavoidable problem since the exploration and

analysis of behavior will always be limited and

insufficient. More concretely, to ensure proper

handling of information, we must consider a semantic

enrichment which extends raw data to other

contextual information representing the meanings.

The main objective of this paper is to propose a

semantic approach for modeling social network users

‘trajectories in order to explore later the behavior of

Internet users. Our approach focuses on solving

problems of modeling data collected from social

networks. Such data are based mainly on related

112

Ounissi, H., Manaa, M. and Akaichi, J.

Towards a Semantic Approach for the Design of Social Network Users’ Geographical Trajectories.

DOI: 10.5220/0006052401120120

In Proceedings of the 8th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2016) - Volume 2: KEOD, pages 112-120

ISBN: 978-989-758-203-5

locations of the user, where we will discuss the

geographical coordinates of locations and also exploit

the proportional metadata. That’s why, we should

make a semantic enrichment of these trajectories to

move from raw trajectories to semantic trajectories by

using domain ontology models that represent the raw

data and their annotations.

The remainder of this paper is organized as

follows: section 2 gives an overview of related work.

Section 3 describes the proposed approach. Section 4

proposes semantic trajectory ontology. Section 5

concludes by stating the main research ideas.

2 RELATED WORK

The following sub-sections report few works in the

literature describing social network trajectory data,

review works on trajectory analysis based on

ontological approaches and introduce the most

popular geospatial ontologies.

2.1 Social Media Trajectory Analysis

The geographical aspect integrated to social networks

has attracted many researchers. Between

(Foursquare, Twitter, Facebook, etc.), numerous

works has been established, where these researchers

have proposed their approaches for achieving a better

understanding of human mobility. Some of them

focus only on behavior analysis, others propose a

semantic enrichment from different sources and

others predict the future locations of users in time and

in space.

On the Foursquare network, a lot of work is

developed. Some of them include the analysis of

behaviors and others focus on the prediction of future

locations. The work of (Kling and Pozdnoukhov,

2012) studied the movement patterns and location

based Foursquare Twitter messages to extract the

semantics behind users’ activities. (Long and al,

2012) explored the check-ins of Foursquare users to

determine their trajectories in order to understand

their preferences sites. Then, the contribution is a

recommendation of appropriate locations for each

intended user. (Preotiuc-Pietro and Cohn, 2013)

offered on the one hand the discovery models of user

behavior according to their movements based on the

clustering techniques. On the other hand, it has

planned their future movements by referring to

successful models. It is a research that is integrated in

order to develop interesting applications such as

recommendation systems. In fact, the authors

undertake to focus on the type of places and their

evolution over time. (Preotiuc-Pietro and Cohn,

2013) have not only sought to go beyond the geo-

coordinates of users over time but also to use

metadata as the name of the location, type, comments

photos. Another study proposed by Krueger and al.

(Krueger and al, 2014) used the social network

Foursquare to enrich mobility data extracted from

another source. For this, they have proposed an

approach to semantically enrich the geospatial data

retrieved from the database e-mobility with the POIs

information provided by the Foursquare service. This

work ensures to guarantee then the behavior analysis

task via the establishment of an interactive

visualization tool.

As to Twitter social networking, similar works has

been done to achieve the same goal. (Fujisaka and al,

2010) analyzed the users' movements based on

Twitter geolocation. They applied a cluster analysis

on frequent POIs. (Cheng and al, 2011) studied the

check-ins to analyze aspects of human mobility:

spatial, temporal, social and textual aspects

associated with his fingerprints. (Ferrari and al, 2011)

have managed to extract urban models from the

location-based social networks (LBSN) in reference

to publications Twitter with reference to their

movements. (Gong and al, 2011) introduced locations

predication system in order to predict the next

location of the user. For this, they used Markov

models. (Ye and al, 2013) have used the information

to model the movement of the user pattern. They

proposed a framework that uses a mixed hidden

Markov model to predict the category of user activity

at the next step and then to predict the most likely

location in view of the distribution of the estimated

category.

The Facebook social network creates also another

search field, where (Backstrom Sun, and Marlow,

2010) have managed to predict the location of such

user by referring to these friends. In fact, research

conducted from this network remains quite limited

because of the high security related to user’s profiles

and both the database will be also quite narrow if we

request to have personal information.

Most of the cited works took trajectories in their

basic states without using a semantic enrichment

unless the examples of (Krueger and al, 2014),

(Preotiuc-Pietro and Cohn, 2013) and (Kling and

Pozdnoukhov, 2012) have tried to add some

additional information to the places while neglecting

the mobile entity that is the user. That’s why

understanding users’ behavior will always remain

average and lack of semantics. Moreover the

techniques used to solve this problem

of understanding human mobility turns around the

Towards a Semantic Approach for the Design of Social Network Users’ Geographical Trajectories

113

techniques of data mining and the markov models.

2.2 Trajectory Analysis based on

Ontological Approaches

Ontologies are used in the context of modeling

trajectories due to their large potential to modelize

and manipulate knowledge and provide a better

understanding for a subject. Furthermore, ontologies

are characterized by semantic interoperability

between individuals, systems or individuals and

systems (Hepp, 2008).

The trajectory modeling problem has been

addressed by various researches. Therefore, we

present a work proposed by Spaccapietra

(Spaccapietra and al, 2008), which defined the notion

of semantic trajectory that are used to decompose the

raw data into a series of decision points i.e., Stop and

Move. This division makes the modelization more

thin and flexible by providing the ability to add other

information coming from different areas.

Thanks to this notion of semantic trajectory,

Baglioni (Baglioni and al, 2008) proposes to increase

the information content of the trajectory through a

semantic enrichment process. In fact, this process

enrichs raw trajectory by adding contextual

information related to the geographic area or

information relevant to the activity of user. Modeling

these trajectories is feasible through the Ontology

Web Language (OWL) formalism. Thereafter, the

obtained ontology covers different components of a

trajectory. Concerning the approach proposed by

authors in (Baglioni and al, 2008), the

conceptualization of a trajectory is represented by a

shutdown sequence connected to a movement. This

connection is made via four relationships: fromStop,

toStop, inMove, outMove. In addition, each Stop is in

a specific temporal dimension as the following

equation: StopHasTime.

In his doctoral thesis (Yan, 2011), Yan offers a

complete architecture for the creation, management

and analysis of trajectories. The architecture is a

modular infrastructure that consists of three main

ontologies: ontology of geometric trajectories,

geographic ontology and domain ontology

application.

The geometric ontology of trajectory is itself

made up of several specialized sub-ontologies in the

description of spatial or temporal concepts.

The geographic ontology contains different

geographical concepts. These concepts can refer to

natural and artificial elements. The ontology of

domain application is specific to the domain in which

integrates the application.

Among the existing work in this context, Perry

(Perry, 2008) chooses to make an analytical

application within the military domain. So he

proposed an ontological approach that reflects theme,

space and time. He is also interested in determining a

set of operators specifically relating to application

requirements. In fact, these operators support the

notion of context and are able to calculate the spatial

and temporal relationships. However, data captured

in this work does not undergo a semantic annotation

process.

In the same spirit, we identify the work of Malki

and al. (Malki and al, 2012) proposed an ontological

approach for the modeling and the analysis of marine

mammals’ trajectories. This trajectory ontology

generated from other ontologies. This approach

considers three separate ontology models: a model of

domain general trajectory, knowledge domain or

semantic model, and time domain model. Other work

is involved in the same axis is that of (Vandecasteele,

2013). In fact, this work took its starting point from

the research done by (Spaccapietra and al, 2008)

speaking about the semantic trajectory and the

research done by (Yan, 2011) proposing to regroup

information following a consisting set of three

ontologies. Vandecasteele (Vandecasteele, 2013)

made an analysis of abnormal behavior of ships to

arrive to control such a maritime surveillance system.

(Hu and al, 2013) proposed a geo-ontology design

pattern for semantic trajectories. The formalization of

this pattern using OWL and they tried to apply the

pattern to two different scenarios, personal travel and

wildlife monitoring.

Recently (Manaa and Akaichi, 2016) proposed a

generic ontology-based for Semantic Trajectory Data

Warehouses using the best of both Data Warehousing

and Data Modeling Semantic worlds.

The choice of using ontological approaches is not

arbitrary. Indeed, the authors sought the best way to

model and to manipulate the trajectories of an

expressive and comprehensive manner to ensure

thereafter the analysis task.

As our work focuses on trajectory modeling, we

are then incited to seek how we can resolve this issue

knowing that ontologies are used both to describe

geographic concepts and to describe spatial concepts.

2.3 Geospatial Ontologies

The ontological modeling studies of the spatial

dimension are quite numerous noting the works of

KEOD 2016 - 8th International Conference on Knowledge Engineering and Ontology Development

114

(INSEE

, COST UCE

, GIEA

, Ordinance Survey

etc.) that are dedicated to the use, the representation

formalism and philosophical rigor employed.

However, works that have full support the description

of geographic features are fairly limited. For this

reason, we focus only to present (FAO

, SUMO

and

BFO

), which are defined as high-level ontologies.

FAO

described the ontology proposed by the

Food and Agriculture Organization of the United

States. This latter was are developed to facilitate the

exchange and share of data in a standardized way

between information systems relating to countries

and/or regions. This ontology manages terms into

multiple languages.

Besides, it has standardized coding systems

(OUN, ISO

, AGROVOC

, etc.). FAO

expresses the

relationship between territories and follows the

historic changes. It was benefited from the

GeoNames

ontology through the addition of

different concepts.

Suggested Upper Merged Ontology (SUMO

)

(Niles and Peace, 2001) is an upper ontology carried

by the IEEE Standard Upper Ontology group (SUO).

The purpose assigned to SUMO

is to establish a

standard able to create a semantic interoperability

between information systems. Moreover, SUMO

has

been developed to describe the geographical area. It

includes hundreds of concepts and relationships. It is

used to design the basic structures of the real world.

Indeed, the description of the real world follows two

trends: type of objects SNAP and type of objects

SPAN.

Moreover, the representation and the

manipulation of geographic features require the

presence of other standard then OWL because

unfortunately it lacks the syntax to describe any

spatial type i.e., point, line and polygon. To fill this

gap, we call the GeoRSS

that is defined as a W3C

Recommendation consortium to describe properties

of geospatial web resources. Besides, GeoRSS

rests

in its most advanced version of a standard called

Geography Markup Language (GML).

GeoRSS

has the advantage of adding the

geographical dimension to the Really Simple

Syndication (RSS) flux, and it was adopted by W3C

as a vocabulary reference for the description of

geospatial properties of web resources. In fact,

http://www.insee.fr/fr/nom_def_met_/xml_rdf.html

www.towntology.net

www.projetgiea.fr

http://www.odnancesurvey.co.uk/oswebsite/ontology/

http://aims.fao.org/geopolitical.owl

Basic Formal Ontology : http://www.iformis.org/bfo

International Organisation for Formalisation -

htttp://www.iso.org/iso/home.html

GeoRSS

is based on the GML standard to be more

generic.

The advantage of using GeoRSS

in the

ontological context is to provide a simplified spatial

representation.

GML is determined by the Open Geospatial

Consortium (OGC

) and respectively used as a

Standard ISO

(ISO19136). It has been defined as a

standardized language based on XML grammar.

GML provides encoding, storage and exchange of the

geometry and the attributes of geographic features

(Cox et al, 2004). Moreover, the GML format ensures

to describe entities, geometry, coordinate systems,

units of measure and symbologies. Each element will

be described in a GML schema that specifies standard

notation and naming.

3 SEMANTIC ANNOTATION OF

TRAJECTORIES OF SOCIAL

NETWORK USERS

A mobile object moves in space and constructs a

trajectory that will itself be formed of a series of

spatio-temporal events i.e., Move and Stop.

Moreover, this spatio-temporal measure seems

insufficient if it is alone retained to carry out the

analysis of an object in movement. Adding additional

information may clarify this movement. In fact, this

additional information will contribute to describe the

context in which a moving object evolves, so the

added value will be a better understanding and

interpretation of movements.

To certify this understanding, we will rely on the

notion of semantic trajectory which was developed

during research works in (Spaccapietra and al, 2008)

and (Baglioni and al, 2008) in order to increase the

rate of information relating to the mobile object via

semantic enrichment process.

This semantic enrichment layer will be from three

perspectives: geographic view, temporal view and

application domain. The result led eventually to

obtain semantic trajectories, easy to be handled and

processed by the user by adopting an expressive and

understandable language. The literature provides in

this context the field of ontologies that are used to

http://www.fao.org/aims/

http://www.geonames.org/ontology/documentation.html

http://www.ontologyportal.org/

http://www.w3.org/2005/Incubator/geo/XGR-geo-

20071023/W3C_XGR_Geo_files/geo_2007.owl

Open Geospatial Consortium – http://www.opengeospatial.org/

Towards a Semantic Approach for the Design of Social Network Users’ Geographical Trajectories

115

meet our needs. Therefore, we use the OWL

formalism to represent these trajectories.

Social networks allow to record the history and

the tracks of daily users. Therefore, a set of metadata

is accumulated through tags, should be useful for the

extraction of the user information related to their

movements.

The following figure (Figure 1) shows a more

concrete example to facilitate modeling trajectory

marked by a well-defined user on a well determined

geo-social network. The user is interested to mark its

current movements and social activities by

identifying his friends and mentioning his mood at

every time. Therefore, we can use its tagging data to

annotate and generate semantic trajectories. The

trajectory below is divided into spatio-temporal

events (Moves and Stops) and enriched by the

contextual data presented on publications.

Stops are

defined with the tags of (Home, University,

Restaurant) and Moves are defined with the tags of

(Train, Path way, Road).

Figure 1: Example of semantic annotation of a trajectory.

4 THE MODELING OF

SEMANTIC CONTEXT

Figure 1 presents a semantic trajectory obtained via

the semantic annotation, using the marking data and

the information on profile. Different trajectories

should be modeled via ontologies in order to achieve

at an early stage the discovering of semantic

behaviors.

To begin our work, we will rely on the ontological

model proposed by Yan (Yan, 2009) (Yan, 2011)

consisting of three ontologies.

 Ontology of trajectory contains different

spatio-temporal concepts necessary for the

geometric description of trajectory.

http://www.w3.org/2006/

 Geographic ontology contains the specific

concepts to the description of a territory.

This ontology will be connected to the

ontology of trajectories and also to the

domain ontology.

 Domain ontology represents the domain to

be studied and the concepts and the needed

relationships.

Our contribution resides in the creation of a

domain ontology related to our subject while we rely

on an existing ontology of trajectory and an existing

geographic ontology. This highlights the interest of

the characteristic of reusability of ontologies.

The model presented in Figure 2 is formed of a:

 Geometric trajectory ontology forms itself

from different ontologies. This ontology

provides the basic concepts to describe the

semantics of trajectory. It consists of three

ontologies:

 Spatio-temporal ontology represented

through the GeoRSS Simple

ontology

to identify space objects.

 OWL Time

ontology to define the

temporal entities.

 Ontology of trajectory that is formed

around the necessary concepts to the

description of a trajectory. We will make

use of the model defined by

(Spaccapietra and al, 2008) to define this

ontology.

 Geographic ontology specifies the

geographical concepts related to

understanding a trajectory consists of two

ontologies FAO

and GeoNames

 Domain ontology will be detailed with

SNDO.

4.1 Social Network Domain Ontology

(SNDO)

We opt for the classification of ontology models

identified by Guarino (Guarino, 1998) which defined,

among other things, the domain ontology. In addition,

our work is based on the life cycle model of an

ontology proposed by Lopez (Lopez, 1999), which is

summed up in three main activities (management,

construction and support). We emphasize the major

classes forming our ontology. Furthermore, in this

context, we specifically interested in the

construction activity respectively, of four stages

(specification, conceptualization, formalization and

KEOD 2016 - 8th International Conference on Knowledge Engineering and Ontology Development

116

Figure 2: Ontological model for semantic trajectories of social network users.

implementation), we will rely on the methodology

proposed by Gandon (Gandon, 2002) to draw up the

tables presenting the concepts, attributes and relations

between concepts.

User profile: This class focuses on presenting

various personal information from the user profile.

This information is defined by social networks such

as Facebook and Twitter. It is divided into five

principal sub-classes (Figure 3).

Figure 3: An excerpt of the design of user profile class.

Social activity: Today social networks are able to

share social activity marked by users. We can then

define these activities following places mentioned on

such publication. In fact, we consider them such as

semantic social activities (Figure 4).

Figure 4: An excerpt of the design of semantic activity

class.

Figure 5: An excerpt of the design of humor class.

Towards a Semantic Approach for the Design of Social Network Users’ Geographical Trajectories

117

Humor: Humors are here, the expression of

user’s feelings, identified during a marking or a

publication. It helps us later to analyze the behavior

of different users (Figure 5).

4.2 Conceptualization

Once

we defined major classes of our domain

ontology, it’s necessary at this stage to extract

concepts and relations and to mention the logical

constraints and inferential knowledge.

 The construction of ontology schema

Ontology is built around multiple concepts connected

by hierarchical relationships, type of specification,

generalization and semantic relations.

The table of concepts is represented in four columns :

the name of the concept (term), the ID is the label of

concept (ConceptID), the ID of parent concept which

the concept is related (ParentID) and the definition of

concept in natural language, as the following table

shows (Table 1).

Ontology concepts are formalized through OWL

and RDF languages. The following example (Figure

6) translates the concepts (user profile, General

information and coordinates).

Figure 6: Concepts (User profile, General information and

coordinates).

 The creation of relations

There are two types of relations: attributes and

associative relationships. Indeed, attributes are the

relation that links a concept to a data type i.e., integer,

string, and date. These attributes have a role to add

knowledge to the concepts themselves without

considering the relationships between these concepts.

This knowledge is determined by the extraction of

information from Twitter social network users.

Table 1: An excerpt of ontology concepts.

Term ConceptID ParentID Definition

General

information and

coordinates

General_Information_and_

coordinates

User Profile Exposes the user information

Humor Humor Thing The mood proven and

mentioned in publications

Date of birth Date_of_birth General_information_and_coordi

nates

Describes the user ‘date of

birth

Table 2: Attributes.

Relation ID attribute Associated concept Attribute type Definition

Date of birth Has_a_date_of_birth User_profile Date Time The date of birth of

each user

Mail Has_a_mail User_profile String Each user has their

own mail

Etudié à Has_studied_in User_profile String School name where

the user continues his

studies.

Table 3: Relations.

Relation ID relation Original concept Target concept Definition

Has identified in Has_identified_in User_profile Publication Indicates that this

profile shares the

same publication

Has a friend Has_a_friend User_Profile User_profile Described the

existence of a

relationship between

two users profiles

Has marked Has_marked User_profile Humor The user express his

mood

KEOD 2016 - 8th International Conference on Knowledge Engineering and Ontology Development

118

The attribute table consists of five columns: the

name of the relation (relation), the id attribute (ID

attribute), the concept which it is attached (associated

concept), the type of the attribute (Attribute Type)

and the attribute definition in natural language

(description) (Table 2).

Associative relations describe the second type.

They connect concepts by relations and links of

subsumption. These relations are frequently binary

and can have different properties example (functional

relationship, transitive relation, etc.)

Table 3 shows relations and consists of five

columns: the name of the relation (relation), the ID of

the relation (ID Relation), the concepts involved in

the relation (Original Concept and target concept) and

finally definition of the relation in natural language.

5 CONCLUSIONS

For the purpose of supporting applications dealing

with mobility data and requesting further contextual

and semantic information about moving objects and

their geographical trajectories, this paper offered a

generic approach for the modeling of social network

users’ geographical trajectories. The proposed

approach combined three ontologies : Social Network

Domain Ontology, Geometric Trajectory Ontology

and Geographic Ontology using GeoNames

and

FAO

Ontology.

As future work, we are going to adopt the

reasoning mechanisms of OWL-DL to deduce new

information and to detect semantic conflicts and gaps

that can hold between heterogeneous trajectory data

sources. Also, we will adopt an ontological approach

to integrate different trajectory data sources in a

trajectory data warehouse by using the shared

trajectory ontology. This integration will support

trajectory-oriented applications dealing with mobility

data, enhance the decision making process and allow

querying mobility data on the semantic level,

revealing then information about the mobile objects

activities and behavior.

REFERENCES

Backstrom, L., Sun, E., and Marlow, C., 2010. Find me if

you can: improving geographical prediction with social

and spatial proximity. In WWW ’10.

Baglioni, M., Macedo, J., Renso, C., Wachowicz, M., 2008.

An Ontology-Based Approach for the Semantic

Modelling and Reasoning on Trajectories. In:

Advances in Conceptual Modeling-Challenges and

Opportunities, Lecture Notes in Computer Science.

Springer Berlin / Heidelberg, pp. 344–353.

Cheng, Z., Caverlee, J., Lee, K., & Sui, D., 2011. Exploring

Millions of Footprints in Location Sharing Services.

In: Fifth International AAAI Conference on Weblogs

and Social Media. Barcelona, Spain.

Cox, S., Daisez, P., Portele, C., Whiteside, A., 2004. Open

GIS Geography Markup Langage (GML)

Implementation Specification (No.OGC 03-105r1)

Open Geospatial Consortium.

Ferrari, L., Rosi, A., Mamei, M., Zambonelli, F., 2011.

Extracting Urban Patterns from Location-based Social

Networks. ACM LBSN ’11, November 1, 2011.

Chicago, IL, USA.

Fujisaka, T., Lee, R., Sumiya, K., 2010. Discovery of user

behavior patterns from geo-tagged micro-blogs. In

Proceedings of the 4th International Conference on

Uniquitous Information Management and

Communication, pp 36.

Gandon, F., 2002. Distributed Artificial Intelligence and

Knowledge Management: Ontologies and Multi-agent

systems for corporate semantic web. Thèse, University

of Nice-Sophia Antipolis.

Gong, Y., Li, Y., Jin, D., Su, L., and Zeng, L., 2011. A

location prediction scheme based on social correlation.

In VTC Spring, 1–5. IEEE.

Guarino, N., 1998. Some Ontological principales for

Designing Upper Level Lexical Ressources. In

Proceeding of the First International Conference on

Lexical Ressources and Evolution.

Hepp, M., 2008. Ontologie : State of the Art, Business

Potentiel and Grand Challenges, Ontology

Management. Springer US. Boston, MA, pp.3-22.

Hu, Y., Janowicz, K., Carral, D., Scheider, S., Kuhn, W.,

Berg-Cross, G., Hitzler, P., Dean, M., Kolas, D., 2013.

A Geo-Ontology Design Pattern for Semantic

Trajectories. Proceeding of Computer Science and

Engineering.

Kling, F., Pozdnoukhov, A., 2012. When a city tells a story:

urban topic analysis. In Proceedings of the 20th

International Conference on Advances in Geographic

Information Systems, pp 482–485.

Krueger, R., Thom, D., Ertl, T., 2014. Visual Analysis

of Movement Behavior using Web Data for Context

Enrichment. IEEE Pacific Visualization Symposium.

Long, X., Jin, L., Joshi, J., 2012. Exploring Trajectory-

Driven Local Geographic Topics in Foursquare.

UbiComp ’12, Sep 5-Sep 8, 2012, Pittsburgh, USA.

Lopez, M., 1999. Overview of Methodologies for

Building Ontologies. Proceeding of the IJCAI’99

Workshop on ontologies and Problem Solving

Methods. Stockholm, Suède, pp.4/1, 4/13.

Malki, J., Bouju, A., Mefteh, W., 2012. Une approche

ontologique pour la modélisation et le raisonnement

sur les trajectoires. Prise en compte des règles

métiers, spatiales et temporelles.

Technique et

Science Informatiques, 31(1): 71-96.

Manaa M., Akaichi J., 2016. Ontology-Based Trajectory

Data Warehouse Conceptual Model. In: Proceeding

of the 18

International Conference Big Data

Towards a Semantic Approach for the Design of Social Network Users’ Geographical Trajectories

119

Analytic and Knowledge Discovery, DaWaK 2016.

Porto, Portugal, pp 329-342.

Niles, I., Pease, A., 2001. Towords a Standard Upper

Ontology. In: Proceeding of the International

Conference on Formal Ontology in Information

Systems, FOIS’01. ACM, New York, NY, USA, pp 2-9.

Perry, M., 2008. A framework to support spatial,

temporal and thematic analytics over semantic web

data. Thèse de doctorates, Wright State

University,Dayton, OH, USA.

Preotiuc-Pietro, D., Cohn, T., 2013. Mining user

behaviours: A study of check-in patterns in location

based social networks.

Spaccapietra, S., Parent, C., Damiani, M. L., Macedo, J.

A., Porto, F., Vangenot, C., 2008. A conceptual view

on trajectories. Data and Knowledge Engineering,

65(1), pp 126-146.

Tryfona, N., Pfoser, D., 2005. Data semantics in location-

based services. pp 168 -195.

Vandecasteele, A., 2013. Modélisation ontologique des

connaissances expertes pour l’analyse de

comportements à risque. Application à surveillance

maritime, (Phd), MINES Paris Tech.

Yan, Z., 2011. Semantic Trajectories: Computing and

Understanding Mobility Data. Thèse.

Ye, J., Zhu, Z., Cheng, J., 2013. What’s Your Next Move:

User Activity Predication in Location-Based Social

Networks?.In: Proceedings of the SIAM International

Conference on Data Mining, SDM2013. Texas, USA.

KEOD 2016 - 8th International Conference on Knowledge Engineering and Ontology Development

120