Towards Semantic Web Mining - OSM

steelsquareInternet και Εφαρμογές Web

20 Οκτ 2013 (πριν από 3 χρόνια και 7 μήνες)

77 εμφανίσεις

Towards Semantic Web
Mining

Bettina Berndt

Andreas Hotho

Gerd Stumme

Semantic Web Mining


Combination of Semantic Web and Web
Mining


Improve Web Mining using Semantic
Web


Improve Semantic Web using Web
Mining


Overview


Web Mining


Extracting Semantics from the Web


Exploiting Semantics for Web Mining


Mining the Semantic Web


Closing the Loop


Conclusion/Assessment

Web Mining



Discovers Local and Global Structure


Structured Data


Goals


Improvement of site design


Generate dynamic recommendations


Improve marketing


Main Areas


Web Content Mining


Web Structure Mining


Web Usage Mining


Content Mining


Type of Text Mining


Uses Tags


Detect co
-
occurrences


Event detection


Reconstruction of page content


Relations in a domain


Web Structure Mining


WebPages as a whole


Uses hyperlinks


Identify relevance


Single Pages


Five types of Web Pages


Head Pages


Navigation Pages



Content Pages


Look up Pages


Personal Pages

Web Usage Mining


Request by Visitors


Additional Structure


Unintended Relationships


Web Mining

Disadvantages of Web Mining


Content/Structure


False positives


Unused


Human understandable


Large amount of data


Usage


Usage tracked by urls


General concepts


Multiplicity of events and urls

Relational

Metadata

DAMLPROJ

COOPERATES

-

WITH

URI
-
GST

URI
-
SWMining

COOPERATES

-

WITH

WORKS
-
IN

PROJECT

RESEARCHER

PERSON

TOP

COOPERATES

--

WITH

TITLE

NAME

RESEARCHER

PERSON

Ontology

COOPERATES

--

WITH

Semantic
Web Mining

WWW

-

URI
-
AHO

Andreas
Hotho


cooperateswith
(
X,Y
)


cooperateswith
(
Y,X
)



WORKS
-
IN

WORKS
-
IN

Outline


Web Mining


Extracting Semantics from the Web


Exploiting Semantics for Web Mining


Mining the Semantic Web


Closing the Loop


Conclusion/Assessment

Extracting Semantics


Ontology Learning


Learn structures of Ontologies


Instance Learning


Populates the Ontologies


Extracting Semantics


Ontology Learning


Semi
-
automatic approach


Merging


FCA
-
Merge


TITANIC


Instance Learning


Information Extraction

Outline


Web Mining


Extracting Semantics from the Web


Exploiting Semantics for Web Mining


Mining the Semantic Web


Closing the Loop


Conclusion/Assessment




How can the semantic web help


with web mining ?

Web Content/Structure Mining


Content Mining


Preprocess the input data


Apply heuristics


Creates a cluster


Web Structure Mining


Page Rank


Keyword Analysis


CLEVER

Conceptual Clustering of Emails (and Bookmarks)


using IE and Formal Concept Analysis for supporting navigation and retrieval.

Web Usage Mining


Goal


Better understand user’s tendencies


Problem


Dynamic pages


How to take advantage of this?


Generate queries


Create usage paths


Classification scheme




Advantages


Structured Model


Improve queries


Analyze single pages


Analyze ontologies


Users history


Outline


Web Mining


Extracting Semantics from the Web


Exploiting Semantics for Web Mining


Mining the Semantic Web


Closing the Loop


Conclusion/Assessment




How can web mining help


build the semantic web?

Semantic Web/Structure
Mining


Intertwined


Relational Data Mining


Looks for patterns


Classification, regression, clustering and
associations


Challenges


Scalability


Distributed

Semantic Web Usage Mining


Goal


Requested page = ontology entity


Log files


Advantages


Understand search strategies


Improve navigation design


Personalize

Outline


Web Mining


Extracting Semantics from the Web


Exploiting Semantics for Web Mining


Mining the Semantic Web


Closing the Loop


Conclusion/Assessment

Mining to Learn Ontologies


Establish a concept hierarchy


OntEx


Determine Association rules


Discover combinations of concepts

Mining to Learn Ontologies

Filling the Ontologies

Use Ontology to Mine

Conclusion/Assesment


Semantic Structures in the Web can
help Web mining


Web Mining can build the Semantic
Web


Combine the two together


Different Idea


Combination of Products

Questions?