semantic web application development

22
DANIEL SLAMOWITZ FACULTY ADVISOR: DR. KNARIG ARABSHIAN The Social Semantic Search Engine

Upload: daniel-slamowitz

Post on 17-Aug-2015

59 views

Category:

Data & Analytics


4 download

TRANSCRIPT

Page 1: Semantic Web Application Development

DANIEL SLAMOWITZ

FACULTY ADVISOR: DR. KNARIG ARABSHIAN

The Social Semantic Search Engine

Page 2: Semantic Web Application Development

Introduction

• Advancement of technology has changed the way we plan dates• Popular applications like Google Maps, Yelp, and Groupon make it

easy to plan out the details• Ability to find a date based on location, activity, rating, price, deals

offered, reviews etc.

Page 3: Semantic Web Application Development

Problem

Planning the perfect date can take forever!

Page 4: Semantic Web Application Development

Solution

• Everything needed to plan a date in one simple application

• Perform a context-aware search that takes into account users’ preferences and locations

• Eliminate the need for manual searching on many individual applications

Page 5: Semantic Web Application Development

Semantic Web Overview

• “The Semantic Web is an extension of the current web in which information is given well-defined meaning, better enabling computers and people to work in cooperation.” – Tim Berners-Lee

• Relationships between resources can made using Triples, which consist of a subject, predicate, and object.

• An Ontology formally defines these concepts and relationships representing a specific domain of interest.

• Resource Description Languages: RDF, RDFS, OWL

Page 6: Semantic Web Application Development

Method

• Limit scope of application due to development time constraints• Retrieve select Arts & Entertainment activities and Cuisines• Limit location to Uptown West and Downtown East NYC

• Use Semantic Web technologies with custom backend Ontology knowledge base

• Populate the ontology using the Yelp Search API

Page 7: Semantic Web Application Development

Features

• Backend ontology knowledge base that interprets data retrieved from the Yelp Search API

• Ability to query specific objects and find matches that are not only exact but also similar

• Ability to search for objects within different domains in a single query

• Ability to search for objects based on two users’ preferences and locations constructing queries that represent a compromise between these data points

Page 8: Semantic Web Application Development

DateNight vs. The Competition

• More accurate and precise search results• Ontology reasoner engine will allow for understanding of a

users’ intent

• Features that no one else offers• Ability to plan a perfect date in a matter of seconds• Ability to search within two different domains• Ability to construct queries based on two users’ preferences

and locations

Page 9: Semantic Web Application Development

System Design & ImplementationHardware/Software Components:

Eclipse Java IDE, Graphical User Interface, Jena Ontology API, JavaScript Object Notation, Laptop with AMD A10 Processor and 16GB DDR3L-1600 MHz of memory, Ontology, Pellet Reasoner, Protégé, Standard Widget Toolkit, Yelp Search API

Page 10: Semantic Web Application Development

System Design & Implementation(cont.)

Page 11: Semantic Web Application Development

Results

Goal: Find restaurants and arts & entertainment activities that are flexible in terms of cuisine and location

Sample Expected OutputKorean restaurants or similar Asian cuisine restaurants and Parks with a rating range of 3.5-4.5 located not only in Little Italy, but in other Downtown East neighborhoods as well.

Page 12: Semantic Web Application Development

Results (cont.)

YelpOutput for 1st Query

Page 13: Semantic Web Application Development

Results (cont.)

Needed 2 separate searches to find a Park near a Restaurant, and still got back Supermarkets and Parking Garages!

YelpOutputFor 2nd Query

Page 14: Semantic Web Application Development

Results (cont.)

DateNight Input

Everything needed in one simple form!

Page 15: Semantic Web Application Development

Results (cont.)

DateNight Output

Retrieved Asian Style Restaurants and Parks

Page 16: Semantic Web Application Development

Results (cont.)

Goal: Invite a date and find restaurants and arts & entertainment activities based on you and your date’s input with an exact location

Sample Expected OutputFrench or Japanese restaurants and Art Galleries or Museums with a rating range of 4.0-5.0 located in the Upper West Side.

Page 17: Semantic Web Application Development

Results (cont.)

DateNight Input

Page 18: Semantic Web Application Development

Results (cont.)

DateNight Output

Page 19: Semantic Web Application Development

Conclusions

• Only requires one search query

• Semantic Web technologies enable smarter more accurate searches of different domains

• Provides improvements upon standard location-based service search• Actually understands users’ intent

Page 20: Semantic Web Application Development

Conclusions (cont.)

Problems during development• Lack of experience using Semantic Web

technologies• Too much data for Pellet Reasoner to handle

• Pellet is not a commercial grade reasoner

• Project scope originally contained every Cuisine and Arts & Entertainment Type on Yelp, and the 5 boroughs of NYC

• Short time span available• Could have implemented other faster

and more efficient reasoners to accomplish task

Solutions

• Sought Dr. Arabshian’s help for guidance

• Used a smaller sample size of resources in the ontology

• Project scope now limited to select Cuisines, Arts & Entertainment Types, and Locations

• Cut down over 30,000 axioms to 1,182 axioms

• Did not have time to implement Google Maps, Google Calendar, or Facebook API’s

Page 21: Semantic Web Application Development

Future Visions

Ultimate Vision: DateNight will become the go-to date planning application

• Additional Features• Expanded ontology with additional locations, cuisines and arts &

entertainment types• Include ability to query for nightlife or shopping

• Elegant web and mobile versions of the user interface• Google Maps, Google Calendar, and Facebook Integration• Provide meetups of more than two people

Page 22: Semantic Web Application Development

References & Acknowledgements

Foundations of Semantic Web Technologies by Hitzler, Krotzsch, and RudolphSemantic Web for the Working Ontologist by Allemang and Hendler

Thanks to Dr. Arabshian for her support, guidance, and encouragement.

Thanks to the staff of Hofstra University’s Computer Science Department for their support during the development of this project.