Abstract | ||
---|---|---|
Many tweets encourage people to behave in a certain way. For instance, a tweet might say “You can lose weight if you eat only apples.” to encourage readers to eat only apples if readers want to lose weight. For this study, we designate such tweets, which compel users to adopt a behavior, as behavioral facilitation tweets. These tweets do not always provide reliable advice; users should be alerted to unreliable tweets. To accomplish this goal, we first extract tweets that are behavioral facilitation tweets. Our studied behavioral facilitation tweets explicitly include behavior-related advice or encouragement. Furthermore, we target not only behavioral facilitation by the author but also behavioral facilitation by others. As described herein, we propose methods of three types to extract behavioral facilitation tweets automatically from numerous tweets: rule-based, support vector machine (SVM), and long short-term memory (LSTM). We use topics of three types to measure the benefits of our proposed methods, comparing rule-based, SVM, and LSTM-based methods to extract behavioral facilitation tweets. Results demonstrated that the LSTM-based method works best. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/BIGCOMP.2019.8679135 | BigComp |
Keywords | Field | DocType |
Support vector machines,Twitter,Data mining,Feature extraction,Training data,Machine learning,Kernel | Kernel (linear algebra),Training set,Facilitation,Computer science,Support vector machine,Feature extraction,Artificial intelligence,Natural language processing | Conference |
ISSN | ISBN | Citations |
2375-933X | 978-1-5386-7789-6 | 1 |
PageRank | References | Authors |
0.37 | 0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Keiichi Mizuka | 1 | 2 | 1.07 |
Yu Suzuki | 2 | 4 | 4.21 |
Akiyo Nadamoto | 3 | 189 | 34.24 |