Partner
Collection
1 Matching Results
Results open in a new window/tab.
Results:
1 - 1 of
1
Automatic trend detection: Time-biased document clustering
This article presents a novel approach of introducing a weighted temporal feature to bias a topic clustering toward articles in a similar time frame, performed over a set of finance journal abstracts from 1974 to 2020 to demonstrate how time can be emphasized in trend detection. The authors detect trending finance topics that are not identifiable when we use a standard clustering approach with no temporal bias.
Date:
March 2, 2021
Creator:
Behpour, Sahar; Mohammadi, Mohammadmahdi; Albert, Mark; Alam, Zinat S.; Wang, Lingling & Xiao, Ting
System:
The UNT Digital Library