Embed Size (px)
DESCRIPTION
Location sharing services(LSS) like Foursquare, Gowalla and Face- book Places gather information from millions of users who leave trails in loca- tions (i.e. chekins) in the form of micro-posts. These footprints provide a unique opportunity to explore the way in which users engage and perceive a point of interest (POI). A POI is as a human construct which describes information about locations (e.g restaurants, cities). In this work we investigate whether the collec- tive perception of a POI can be used as a real-time dataset from which POI’s transient features can be extracted. We introduce a graph-based model for profil- ing geographical areas based on social awareness streams. Based on this model we define a set of measures that can characterise a location-based social aware- ness stream as well as act as indicators of volatile events occurring at a POI. We applied the model and measures on a dataset consisting of a collection of tweets generated at the city of Sheffield and registered over three week-ends. The model and measures introduced in this paper are relevant for design of future location-based services, real-time emergency-response models, as well as traffic forecasting. Our empirical findings demonstrate that social awareness streams not only can act as an event-sensor but also can enrich the profile of a location-entity.
Citation preview
A.E. Cano, A. Varga and F. Ciravegna The Oak Group, �
Department of Computer Science, �The University of Sheffield
Volatile Classification Volatile Classification of Point of Interests based on Social Awareness Streams
• Introduction • Motivation
• Contributions
• Related Work • Geolattice Awareness Streams
• Transient Semantic Classification of a POI
• Experiments • Conclusions and Future Work
Outline Outline
Introduction Introduction Social Awareness Streams
[1] M. Naaman, J. Boase, and C. H. Lai. Is it really about me?: message content in social awareness streams. In CSCW ’10: Proceedings of the 2010 ACM conference on Computer supported cooperative work, pages 189–192, New York, NY, USA, 2010. ACM.
Collection of semi-public, natural language messages produced by different users and characterised by their brevity
Introduction Introduction Social Awareness Streams
Used as a historical data set for profiling users’ activities and behavior [2][3][4].
Time
Murakami
Michio Kaku
Spotify
Iphone
€
Δt
Contextual Meaning of Places
Introduction Introduction
Introduction Motivation Point of Interest (POI)
o Specific point location that someone may find useful or interesting.
o Characterised by static data (e.g. name, geo-coordinates) o Classified by the type of services it provides.
Coordinates: 51.507221, -0.1275
London, United Kingdom
Categories: Leisure and entertainment Literature, film and television Museums and art galleries Music Sports
Source : http://en.wikipedia.org/wiki/London
Introduction Motivation Place
An evolving set of both communal and personal labels for potentially overlapping geometric volumes.
[5] Hightower, J.: From Position to Place. In: Proc. of the 2003 Workshop on Location-Aware Computing. (2003) 10–12 part of the 2003 Ubiquitous Computing Conf.
A meaningful place can capture the venues’ demographical, environmental, historical, personal or commercial significance.
Introduction Motivation Social Awareness Streams and POIs
Can social awareness streams convey information for inducing a transient classification of a point of interest?
Introduction Motivation Point of Interest (POI)
Tourism
History
Fashion Shopping
Literature
Shopping
History
Literature
Tourism
Fashion
Introduction Motivation Point of Interest (POI)
Tourism
History
Fashion Shopping
Literature
Shopping
History
Literature
Tourism
Fashion
Tags: looting, unrest, police..
Introduction Motivation Point of Interest (POI)
Tourism
History
Fashion Shopping
Literature
Shopping
History
Literature
Tourism
Fashion
Tags: looting, unrest, riots, police..
Social Crisis
Violence
Politics Uprising
Related Work Related Work Use of SAS for Tagging Places
• Cheng et al [6]: Spatial distribution of words in tweets for predicting users’ location. Their probabilistic framework can identify words in tweets with strong geo-scope.
• Cheng et al [7]: Mobility patterns of users in LSS. They correlate social status, geographic and economic factors with mobility and perform sentiment analysis of posts for deriving unobserved context between people and location.
• Lin et al [8]: Taxonomy of ‘place naming methods’, showing that a person’s perceived familiarity with a place and the variety of people who visit it, strongly influence the way people refer to this place when interacting with others.
Related Work Related Work Use of SAS for Tagging Places
• Ireson and Ciravegna [9]: Toponym resolution (i.e. allocation of specific geo-location to target location terms) using Flickr data. Using geo-tag media, users’ contacts, content as features, they build an SVM classifier for learning location labels associated to a location term.
Rather than identifying words for tagging places we study how location-based generated content can be modelled for discovering topics or categories that classify a place in time. For this we present:
GeoLattice Awareness Streams Model Transient Semantic Classification of a POI
Related Work Related Work Contrasting our Work
A Geolattice Awareness Stream is a sequence of tuples
S:= (Poiq1,Mq2,Rq3,Y,ft)
GeoLattice Aware GeoLattice Awareness Streams
Definition 1.
A Geolattice Awareness Stream is a sequence of tuples
S:= (Poiq1,Mq2,Rq3,Y,ft)
GeoLattice Aware GeoLattice Awareness Streams
Definition 1.
Name Geographical-bounding area Geo-coordinates
A Geolattice Awareness Stream is a sequence of tuples
S:= (Poiq1,Mq2,Rq3,Y,ft)
GeoLattice Aware GeoLattice Awareness Streams
Definition 1.
source
A Geolattice Awareness Stream is a sequence of tuples
S:= (Poiq1,Mq2,Rq3,Y,ft)
GeoLattice Aware GeoLattice Awareness Streams
Definition 1.
Keywords Categories Hashtags
Given a GeoLattice Awareness Stream, a POI awareness stream can be defined as the sequence of tuples from S where:
GeoLattice Aware GeoLattice Awareness Streams
POI Awareness Stream
GeoLattice Aware GeoLattice Awareness Streams
Steel
Education
Football
Industrial Revolution
Gothic Revival
Modern Architecture
Supertram Urban Zone
Archeology
Bath
Politics
Thames
Westminster
POI Awareness Stream
UK
Transient Semant Transient Semantic Classification of a POI
How can we derive a temporal classification (Rcat) of a POI given it’s POI Awareness Stream (S(Poi’)) ?
Problem
Transient Semant Transient Semantic Classification of a POI
Approach
Retrieve Messages from a POI Awareness Stream for a window of time [ts-te]
Transient Semant Transient Semantic Classification of a POI
Approach- Message Enrichment
Retrieve Messages from a POI Awareness Stream for a window of time [ts-te]
Enrich Messages
Transient Semant Transient Semantic Classification of a POI
Approach- Message Enrichment
Retrieve Messages from a POI Awareness Stream for a window of time [ts-te]
Enrich Messages
Transient Semant Transient Semantic Classification of a POI
Approach- Semantic Categorisation
Retrieve Messages from a POI Awareness Stream for a window of time [ts-te]
Enrich Messages Semantic
Categorisation
Transient Semant Transient Semantic Classification of a POI
Approach- Semantic Categorisation
Retrieve Messages from a POI Awareness Stream for a window of time [ts-te]
Enrich Messages Semantic
Categorisation
Transient Semant Transient Semantic Classification of a POI
Approach- Induce Category Function
Retrieve Messages from a POI Awareness Stream for a window of time [ts-te]
Enrich Messages Semantic
Categorisation
Induce Category Function
Transient Semant Transient Semantic Classification of a POI
Approach- Induce Category Function
Category Entropy of a stream
Indicates the topical diversity of the stream.
Low category entropy levels reveal that a stream is dominated by few categories, while a high category balance reveals a higher topical diversity.
Transient Semant Transient Semantic Classification of a POI
Approach- Induce Category Function
Category Entropy of a stream
Normal Conditions Normal Conditions Broken (Events Happening)
Restaurant
Food
Italian Cuisine CE low
City
CE high
Restaurant
CE Increases
Food
Italian Cuisine
City
CE Decreases
Transient Semant Transient Semantic Classification of a POI
Approach- Induce Category Function
Category Entropy of a stream
Normal Conditions Normal Conditions Broken (Events Happening)
Restaurant
Food
Italian Cuisine CE low
City
CE high
Restaurant
CE Increases
Food
Italian Cuisine
City
CE Decreases
Transient Semant Transient Semantic Classification of a POI
Approach- Induce Category Function
Mutual Information
Measures the information that two discrete random variables share. In this work we consider the MI between:
Categories- Hashtags(MI)
Hashtags- Keywords(MI)
Categories- Keywords(MI)
Transient Semant Transient Semantic Classification of a POI
Approach- Induce Category Function
Mutual Information
Measures the information that two discrete random variables share. In this work we consider the MI between:
Categories- Hashtags(MI)
Hashtags- Keywords(MI)
Categories- Keywords(MI)
Enpirical Study Empirical Study
Data Set
We monitored tweets generated from the city of Sheffield, during 3 weekends.
Common – 2011-06-10 to 2011-06-13 Food Festival – 2011-07-08 to 2011-07-11 Tramlines Music Festival – 2011-07-22 to 2011-07-25
Enpirical Study Empirical Study
Data Set
After enriching the content of the tweets and querying DBpedia we got the following hashtags, resources, and categories:
Enpirical Study Empirical Study
Data Set
The ten top more frequent hashtags in the data sets were :
Results Results
Category Entropy
Category entropy for each category of each of the three data set.
Results Results
Category Entropy
Category entropy for each category of each of the three data set.
Results Results
Categories with lowest Category Entropy
Results Results
Categories with lowest Category Entropy
Results Results
Categories with lowest Category Entropy
Event involving 2012 Olympic Tickets sales
Event involving Spanish Football
Results Results
Categories with lowest Category Entropy
Results Results
Deriving Context with MI
To provide a context in which the a category is used, we used the mutual information between categories and hashtags, from which we obtained a set of hashtags used to further derive related keywords (using Hashtag-Keywords MI).
Results Results
Highlights
A point of interest considered as a Location-Entity presents the “meformer” and “informer” patterns observed by Naaman et al.[9] in Person-Entity activity streams.
Location-Entity Meformer pattern refers to a self focus of a Location-Entity, presenting information about endemic events.
Location-Entity Informer pattern refers to an information sharing about external events, not necessarily related to this Location-Entity.
• A formalisation for describing geographically bounded social awareness streams.
• An approach to derive transient categorisations of Points of Interest based on DBPedia Categories.
Conclusions Conclusions and Future Work
We have presented :
Future Work:
• Determining a semantic coherence metric which could induce broader category clusters.
References References
[1] M. Naaman, J. Boase, and C. H. Lai. Is it really about me?: message content in social awareness streams. In CSCW ’10: Proceedings of the 2010 ACM conference on Computer supported cooperative work, pages 189–192, New York, NY, USA, 2010. ACM. [2] C. Wagner and M. Strohmaier. The wisdom in tweetonomies: Acquiring latent conceptual structures from social awareness streams. In Proc. of the Semantic Search 2010 Workshop (SemSearch2010), april 2010. [3] F.Abel,Q.Gao,G.Houben,andK.Tao.Semanticenrichmentoftwitterpostsforuserprofile construction on the social web. In In proceedings of Extended Semantic Web Conference 2011, May 2011. [4] A. Cano, S. Tucker, and F. Ciravegna. Capturing entity-based semantics emerging from personal awareness streams. In Proceedings of the Workshop on Making Sense of Microposts (MSM2011), May 2011.
References References
[6] Hightower, J.: From Position to Place. In: Proc. of the 2003 Workshop on Location-Aware Computing. (2003) 10–12 part of the 2003 Ubiquitous Computing Conf. [7] Z. Cheng, J. Caverlee, and K. Lee. You are where you tweet: a content-based approach to geo-locating twitter users. In Proceedings of the 19th ACM international conference on Information and knowledge management, CIKM ’10, pages 759–768, New York, NY, USA, 2010. ACM. [8] K. L. D. Z. S. Zhiyuan Cheng, James Caverlee. Exploring millions of footprints in locationsharing services. In 5th International Conference on Weblogs and Social Media (ICWSM), ICWSM ’11. ACM, 2011. [9] J. Lin, G. Xiang, J. I. Hong, and N. Sadeh. Modeling people’s place naming preferences inlocation sharing. In Proceedings of the 12th ACM international conference on Ubiquitous computing, Ubicomp ’10, pages 75–84, New York, NY, USA, 2010. ACM. [10] N. Ireson and F. Ciravegna. Toponym resolution in social media. In Proc., 9th InternationalSemantic Web Conference. ISWC 2010, 2010.
![[Japanese] How Reactive Streams and Akka Streams change the JVM Ecosystem @ Reactive Shinjuku](https://img.pdfslide.tips/doc/110x75/58ef690e1a28abf8088b45e5/japanese-how-reactive-streams-and-akka-streams-change-the-jvm-ecosystem-.jpg)
