Abstract
Online Social networks are widely used in current times. In this paper, we investigate twitter posts to identify features that feed in intention mining calculation. The posts features are divided into three sets: tweets textual features, users features, and network contextual features. In this paper, our focus is on tweets analysing textual features. As a result of this paper, we were able to create intentions profiles for 2960 users based on textual features. The prediction accuracy of three classifiers was compared for the data set, using ten intention categories to test the features. The best accuracy was achieved for SVM classifier. In the future, we plan to include user and network contextual features aiming at improving the prediction accuracy.
Original language | English |
---|---|
Title of host publication | Intelligent Data Engineering and Automated Learning – IDEAL 2019 - 20th International Conference, Proceedings |
Editors | Hujun Yin, Richard Allmendinger, David Camacho, Peter Tino, Antonio J. Tallón-Ballesteros, Ronaldo Menezes |
Publisher | Springer |
Pages | 121-128 |
Number of pages | 8 |
ISBN (Print) | 9783030336066 |
DOIs | |
Publication status | Published - 2019 |
Externally published | Yes |
Event | 20th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2019 - Manchester, United Kingdom Duration: 14 Nov 2019 → 16 Nov 2019 |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 11871 LNCS |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 20th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2019 |
---|---|
Country/Territory | United Kingdom |
City | Manchester |
Period | 14/11/19 → 16/11/19 |
Bibliographical note
Publisher Copyright:© 2019, Springer Nature Switzerland AG.
Keywords
- Feature selection
- Intention mining
- Machine learning
- Online Social Network