원문정보
초록
영어
As the amount of cellphone text message use has increased, spam text messages also have increased. Presently in mobile devices, spam filtering methods are in a very basic level such as simple character string comparison or specific number blocking. Typical filtering methods such as bayesian classifier, logistic regression and decision tree for detecting spam messages take quite a long time. In order to perform spam filtering with these methods, high performance computer resources and lots of SMS samples are required. In addition, if servers come to store normal messages, the problem of personal information infringement could arise. For mobile devices to independently perform spam filtering, there are many limitations in the aspects of storage space, memory, and CPU processing capability. Thus, this study tries to propose light and quick algorithm through which SMS filtering can be performed within mobile devices independently.
목차
1. Introduction
2. Related Work
2.1. SMS Spam Collection v.1 Data Set
2.2. Data Mining Algorithm
3. Experimental Study
3.1. Experimental Setup
3.2. Proposed Method
3.3. SMS Spam Collection Data Set
3.4. Selected Variables
4. Discussion
Acknowledgements
References