SMS spam detection using selected text features and Boosting Classifiers
Authors:
- 1 May 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
Short Message Service (also called SMS), is the product of modern mobile communication, which provides more convenience and options for communication. The unwanted advertisements sent with SMS significantly interferes the normal communication among costumers, which should have been guaranteed by the mobile service providers. In this paper, we propose a GentleBoost algorithm for SMS spam detection. Because we have unbalanced data and for this type of data, Boosting Classifiers are good choice. Also, we found that in most cases, especially when we have binary Classification and unbalanced data, GentleBoost performs better than other Boosting Classifiers and leads to better performance. Therefore, we choose GentleBoost as preferred Classifier, also we guarantee that no one use this technique for SMS spam detection and this method is a completely novel method. Another important advantage of the suggested approach is related to the way of extracting the word attributes because by removing unused features and optimizing them, we reduced the number of word attributes significantly without reducing accuracy. Therefore, Our propose approach obtained high accuracy with minimum storage consumption.Keywords
This publication has 8 references indexed in Scilit:
- Detection of Cross Site Scripting Attack in Wireless Networks Using n-Gram and SVMMobile Information Systems, 2012
- Intelligent spam classification for mobile text message2011
- Independent and Personal SMS Spam Filtering2011
- SMSAssassin2011
- Hybrid spam filtering for mobile communicationComputers & Security, 2010
- An Anti-SMS-Spam Using CAPTCHA2008
- Multiclass text categorization for automated survey coding2003
- Additive logistic regression: a statistical view of boosting (With discussion and a rejoinder by the authors)The Annals of Statistics, 2000