Predicting the influence of microblog entries regarding public health emergencies

An, Lu; Yi, Xingyue; Yu, Chuanming; Li, Gang

Permalink

https://hdl.handle.net/2142/96686 Copy

Description

Title

Predicting the influence of microblog entries regarding public health emergencies

Author(s)

• An, Lu
• Yi, Xingyue
• Yu, Chuanming
• Li, Gang

Issue Date

2017

Keyword(s)

• Public health emergencies
• Influence of microblog entries
• Prediction
• Random forest
• BM25 Latent Dirichlet Allocation model (LDA-BM25)

Abstract

Predicting the influence of microblog entries regarding public health emergencies can help management departments improve the prospectiveness of decision making. In this study, we measure the influence of microblog entries regarding public health emergencies from their forwarding, comment and favorite counts. A microblog influence prediction model, which is comprised of user, time and content features, is proposed by using the random forest method and the BM25 Latent Dirichlet Allocation model (LDA-BM25). Microblog entries on the Ebola outbreak are selected as test data. Results reveal that the proposed model can accurately predict the influence of microblog entries regarding public health emergencies with the accuracy rate reaching 88.8%. Individual features, which play a role in the influence of microblog entries, and their influence inclination are also analyzed. The findings of the study can help management departments of public health emergencies predict the upcoming salient issues, and take appropriate measures in advance.

Publisher

iSchools

Series/Report Name or Number

iConference 2017 Proceedings

Type of Resource

text

Language

en

Permalink

http://hdl.handle.net/2142/96686

Copyright and License Information

Copyright 2017 Lu An, Xingyue Yi, Chuanming Yu, and Gang Li

Owning Collections