Deep Dives into TopBraid EVN, Part 1: Automated Tagging with the New AutoClassifier October 15, 2015

Transcription

1 Deep Dives into TopBraid EVN, Part 1: Automated Tagging with the New AutoClassifier October 15, 2015 Copyright 2015 TopQuadrant Inc. Slide 1

2 Today s Program I. What can the AutoClassifier do for you? II. AutoClassifier in Action III. Questions and Answers Bob DuCharme Richard Cyganiak Copyright 2015 TopQuadrant Inc. Slide 2

3 Our Mission: TopQuadrant Company Empower people by making enterprise information meaningful Our Foundation: TopQuadrant was founded on and continues a strong commitment to Semantic Web standards Our Evolution: Tools/Platform Services Company (2002) Company (2006) Business Solution Company ( Today) Copyright 2015 TopQuadrant Inc. Slide 3

4 TopBraid Enterprise Vocabulary Net (TopBraid EVN) TopBraid EVN: A flexible web-based solution for managing taxonomies and business vocabularies in support of business objectives for: Content delivery Search Navigation Data / metadata integration Disambiguation of terms Copyright 2015 TopQuadrant Inc. Slide 4

5 TopBraid EVN Capabilities include: Vocabulary Editing: Cloning, merging, repositioning and re-numbering of concepts; global edit operations for selected groups of concepts Easy Customization: Easy customization of user interfaces, custom validation rules and processing via SPARQL Rules and SPARQLMotion Reporting: Built-in reports as well as advanced query-building tools and reporting through graphical interfaces Merging, Importing, and Exporting: From and to RDBMs, spreadsheets (CSV & Excel), XML, RSS feeds, SPARQL endpoints, RDF and OWL Systems Integration: Integrates with existing enterprise or vocabulary management systems via Web Services interfaces and APIs Enterprise-ready: Scalable architecture offers a choice of relational and RDF database repositories with LDAP integration for access control Unlimited Work-in-Progress Copies: Virtual work-in-progress copies of vocabularies enable controlled publishing, review and approval Copyright 2015 TopQuadrant Inc. Slide 5

6 Curating a taxonomy with EVN Copyright 2015 TopQuadrant Inc. Slide 6

7 Using a taxonomy with Tagger Copyright 2015 TopQuadrant Inc. Slide 7

8 AutoClassifier Demo Copyright 2015 TopQuadrant Inc. Slide 8

9 How good are the results? Source of original slides: Alyona Medelyan, Text Analytics on 2 Million Documents: A Case Study

10 Professional indexers keywords* Agrovoc terms: dietary guidelines overweight meal patterns diet weight reduction energy value feeding habits al disorders al requirements al physiology nutrient excesses status public health disease control regulations programs surveillance food developing countries developed countries price formation foods urbanization overeating food intake food consumption globalization human fiscal taxes price direct taxation prices * 6 professional FAO indexers assigned terms from the Agrovoc thesaurus to the same document, entitled The global obesity problem

11 Agrovoc terms: overweight overeating feeding habits al requirements human surveillance al physiology status food foods food intake food consumption dietary guidelines diet nutrient excesses meal patterns programs public health disease control al disorders price fiscal prices taxes developing countries globalization weight reduction urbanization developed countries price formation energy value direct taxation regulations Indexer 1: Indexer 2: Comparison of 2 indexers

12 Comparison of CS students* & Maui * 15 teams of 2 students each assigned keywords to the same document, entitled A safe, efficient regression test selection technique

13 Human vs. algorithm consistency 15 teams of 2 CS students vs. Maui on 20 CS documents & Wikipedia voc Method Min Avg Max Students Maui

14 AutoClassifier benefits Add value to more content faster and at a larger scale Which adds value to applications that use that content Makes it easier for customers and employees to find that content Inexpensively, because of the scale Tag all documents or a subset Machine learning no need to manually create rules Review process lets humans have the last word Integrates with other systems using APIs Copyright 2015 TopQuadrant Inc. Slide 14

15 Thank you! Request EVN evaluation demo server: Copyright 2015 TopQuadrant Inc. Slide 15