Sentimental analysis of Chinese new social media for stock market information

Guanhang Chen, Lilin He, Konstantinos Papangelis

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

1 Citation (Scopus)

Abstract

The popularity of social media provides a new platform to collect big social data. With the development of social sentiment analysis, high business value extracted from social data are applied to various fields. Asset price prediction, as an emerging topic based on the behavioral economics, is closely linked to social data analysis. This research aims to explore the effort of sentiment analysis data in the prediction of China composite index. Data from Sina Weibo and financial community is processed to get the useful sentiment information. A linear regression model and a multilayer neural network algorithm are used to prove the relationship between social data and price market prediction. The experiments show a strong relationship between the numbers of negative sentiment and a multilayer perceptron model is effectively built to predict the composite index.

Original languageEnglish
Title of host publicationPRAI 2019 - Proceedings of 2019 International Conference on Pattern Recognition and Artificial Intelligence
PublisherAssociation for Computing Machinery
Pages1-6
Number of pages6
ISBN (Electronic)9781450372312
DOIs
Publication statusPublished - 26 Aug 2019
Event2019 International Conference on Pattern Recognition and Artificial Intelligence, PRAI 2019 - Zhejiang, China
Duration: 26 Aug 201928 Aug 2019

Publication series

NameACM International Conference Proceeding Series

Conference

Conference2019 International Conference on Pattern Recognition and Artificial Intelligence, PRAI 2019
Country/TerritoryChina
CityZhejiang
Period26/08/1928/08/19

Keywords

  • Sentiment analysis
  • stock market

Fingerprint

Dive into the research topics of 'Sentimental analysis of Chinese new social media for stock market information'. Together they form a unique fingerprint.

Cite this