Maciej Janik Outline Document categorization

Maciej Janik

Outline

Document categorization

Document categorization by people features, graph ontology category definition categorization context

Automatic text categorization

Classic categorization

Graph representation of text

Graph representations examples

Graph-based categorization

Ontology

Ontology - example

Use of ontologies in classification

Ontology-based classification

Onto. Categorization – bases

Classes and categories

Who? What? Where? When? Why?

Onto. Categorization system

Algorithm sketch

Algorithm sketch – more details

Experiments

Experiments (2)

Algorithm details (1)

Example of entity matching Ford Motor Company Jaguar (animal) Jaguar Cars Ltd. Chief Executive Officer Process (computing) Business process Process (science) Land_Rover Alan_Mulally Sales Ford Motor Company

Algorithm details (2)

Example – parse tree and triples

Algorithm details (3)

Algorithm details (4)

Thematic graph examples

Algorithm details (5)

Algorithm details (6)

Entities and categories

Longer example

Longer example

Longer example graph properties

Longer example categories

Wikipedia categories

Overall performance of initial tests

Text corpora – CNN news

CNN and Wikipedia

Categorization results - BOW

Categorization results Wikipedia

Summary of work

Remaining research

Current work in progress

Proposed work

Published papers

References