Date of Graduation
Bachelor of Science
Computer Science and Computer Engineering
Committee Member/Second Reader
This paper explores the value of information contained in spam tweets as it pertains to prediction accuracy. As a case study, tweets discussing Bitcoin were collected and used to predict the rise and fall of Bitcoin value. Precision of prediction both with and without spam tweets, as identified by a naive Bayesian spam filter, were measured. Results showed a minor increase in accuracy when spam tweets were included, indicating that spam messages likely contain information valuable for prediction of market fluctuations.
Anderson, S. C. (2015). The Information of Spam. Computer Science and Computer Engineering Undergraduate Honors Theses Retrieved from https://scholarworks.uark.edu/csceuht/34