DETERMINATION OF TEXT RELEVANCY BASED ON KEYWORDS ASSOCIATION FOR INTERACTIVE NEWS NETWORK
| Main Author: | S.M.F.D Syed Mustapha |
|---|---|
| Format: | Article |
| Terbitan: |
, 2019
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/3491731 |
Daftar Isi:
- News network is an initiative to allow the content of several news to be associated by the contextual information such as event, people and location. This information may use common and proper noun to describe the similar context of an object, such as “Washington D.C” and “the capital city” or “former president” and “Barrack Obama”. These words cooccur in various times such that they can be associated as keywords to describe certain context on the content of the news. In the literature, many approaches and techniques on the keywords extraction have been discussed but it is argued the lacking on keywords association based on the context of the text, particularly news. Associated keywords are used to “synonymize” the words that are related by the context of news rather than merely observing syntactical or synonymical values. From these words association, news network can be built from the news corpus such that news structure is a stratification that is based on its relevancy to the set of keywords. Named entity recognizer which is a known research area plays significant role in characterizing the relationship in the news network such that the relevancy between news are understood semantically. Event contains the essence of the news that is made up of the activities, actors who are involved in the activities, location and other non-living objects that made up part of the event, called signatures. The results demonstrate the formation of associated keywords based on the context and the building of the news network.