Mining Complex Entities from Heterogeneous Information Networks
author: Fabio Ciravegna, Department of Computer Science, University of Sheffield
author: Andrea Varga, Department of Computer Science, University of Sheffield
published: Nov. 29, 2011, recorded: September 2011, views: 5194
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Most research on information mining has focused on classic Information Extraction (IE) tasks, from structured and unstructured documents, like newspaper articles and web pages. In the last years however the staggering growth of social media as platform for sharing content has moved the focus towards a different type of extraction target. Social media pose a number of challenge to information extraction: contributions to social media sites like blogs, forums, Twitter, etc. are conversational in nature and thus tend to be brief and informal, containing imprecise, subjective and ambiguous information. The expanded context (who the author is, the social and geographical context, their social links, etc.) becomes relevant to disambiguate and interlink information.
Aim of this tutorial is to introduce and discuss issues, methodologies and technologies for extracting information from documents, with a particular focus on mining heterogeneous information networks (e.g. social websites) in order to mine complex entities.
The tutorial covers:
- Introduction to information extraction from documents in general (20 minutes) and from information networks in particular (10 minutes)
- Introduction to machine learning based methods for information extraction (75 minutes)
- representing documents and feature sets
- entity and terminology recognition
- learning gazetteers
- event and relation extraction
- extraction from multimedia documents
- Annotation for training (15 minutes)
- feature selection
- annotation and error
- porting across domains
- Information Extraction from information networks (45 minutes)
- using the Twitter and Facebook APIs
- entity recognition and resolution
- term association
- entity disambiguation over large scale
- Conclusion and future work (15 minutes)
The focus is on Machine Learning based methods. We will cover - among others - methods using Rule Induction, SVM, CRF, HMM, Transfer Learning, Active Learning. We will Also discuss real world cases from the field of information and knowledge management.
Download slides: ecmlpkdd2011_ciravegna_varga_networks.pdf (9.5 MB)
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !
Write your own review or comment: