Difference between revisions of "Is this the same TV program?"
Line 1: | Line 1: | ||
{{Projectproposal | {{Projectproposal | ||
|Contact person=Valentina Maccatrozzo | |Contact person=Valentina Maccatrozzo | ||
− | |Master areas= | + | |Master areas=Internet and Web Technology, Multimedia, Information and Communication Technology, Computer Science and Communication, Information Sciences |
|Project page=http://vista-tv.eu/ | |Project page=http://vista-tv.eu/ | ||
− | |Fulfilled= | + | |Fulfilled=Yes |
}} | }} | ||
+ | |||
+ | |free text=}} | ||
One of the goals of the ViSTA-TV project is to present users with live recommendations for TV programs they may be interested in. To improve the current recommendation system, we are looking for external information sources in order to 'enrich' standard EPGs (Electronic Program Guides). The problem with most of these external sources is that they are published as free text, i.e., news paper articles, blog posts, or tweets. The challenge here is to make the computer understand which TV program a piece of text is about. For instance, if we have a news article talking about "the last Batman movie", the date of the article might be an important indicator to identify which Batman movie the article is about. | One of the goals of the ViSTA-TV project is to present users with live recommendations for TV programs they may be interested in. To improve the current recommendation system, we are looking for external information sources in order to 'enrich' standard EPGs (Electronic Program Guides). The problem with most of these external sources is that they are published as free text, i.e., news paper articles, blog posts, or tweets. The challenge here is to make the computer understand which TV program a piece of text is about. For instance, if we have a news article talking about "the last Batman movie", the date of the article might be an important indicator to identify which Batman movie the article is about. | ||
Latest revision as of 12:46, 25 October 2016
Contents
About Is this the same TV program?
- Contact person: has supervisor::Valentina Maccatrozzo
- This project has been fulfilled.
- This project fits in the following Bachelor programs: {{#arraymap:|, |xXx|bachelorproject within::xXx|,}}
- This project fits in the following masterareas: {{#arraymap:Internet and Web Technology, Multimedia, Information and Communication Technology, Computer Science and Communication, Information Sciences|, |xXx|project within::xXx|,}}
- Project website: has projectpage:: http://vista-tv.eu/
Description
|free text=}} One of the goals of the ViSTA-TV project is to present users with live recommendations for TV programs they may be interested in. To improve the current recommendation system, we are looking for external information sources in order to 'enrich' standard EPGs (Electronic Program Guides). The problem with most of these external sources is that they are published as free text, i.e., news paper articles, blog posts, or tweets. The challenge here is to make the computer understand which TV program a piece of text is about. For instance, if we have a news article talking about "the last Batman movie", the date of the article might be an important indicator to identify which Batman movie the article is about.
Your Goal
The objective of this Master Project is identify a methodology that, given a description of a TV program and a piece of text from an external text, it is possible to indicate a) which program the article is about and b) what the probability of confidence of this suggestion is. As there are already many disambiguation technologies that aim at identifying the topic of a piece of text, you will not develop these from scratch but compare them to see which one works best for TV programs.
Tasks
- evaluate and compare topic disambiguation techniques
- develop confidence measure to indicate how certain the system is of its decision
- represent data in a structured way, using unique IDs, removing duplicates etc.
Tools and Data
- Coming soon
Recommended Prior Knowledge
Web-gebaseerde Kennisrepresentatie, Semantic Web, Information Retrieval