Bi-Term Topic Model for SMS Classification

Bi-Term Topic Model for SMS Classification

Jialin Ma (The Laboratory for Internet of Things and Mobile Internet Technology of Jiangsu Province, Huaiyin Institute of Technology, Huaian, China & College of Computer and Information, Hohai University, Nanjing, China), Yongjun Zhang (Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian, China & College of Computer and Information, Hohai University, Nanjing, China), Lin Zhang (Faculty of Management Engineering, Huaiyin Institute of Technology, Huaian, China), Kun Yu (Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian, China) and Jinlin Liu (Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huaian, China)
DOI: 10.4018/ijbdcn.2017070103
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

With the overflowing of Short Message Service (SMS) spam nowadays, many traditional text classification algorithms are used for SMS spam filtering. Nevertheless, because the content of SMS spam messages are miscellaneous and distinct from general text files, such as more shorter, usually including mass of abbreviations, symbols, variant words and distort or deform sentences, the traditional classifiers aren't fit for the task of SMS spam filtering. In this paper, the authors propose a Short Message Biterm Topic Model (SM-BTM) which can be used to automatically learn latent semantic features from SMS spam corpus for the task of SMS spam filtering. The SM-BTM is based on the probability of topic model theory and Biterm Topic Model (BTM). The experiments in this work show the proposed model SM-BTM can acquire higher quality of topic features than the original BTM, and is more suitable for identifying the miscellaneous SMS spam.
Article Preview

1. Introduction

Short Message Service(SMS) spam messages are unsolicited and unwanted messages. With the popularization of mobile phone, especially the smart phone in last ten years, SMS plays important roles on daily basis. Therefore, SMS contains huge commercial and advertising values. At the same time, criminals pay attention to abuse SMS to defraud or send various kinds of advertisements (Almeida, Hidalgo, & Yamakami, 2011; Sohn, Lee, & Rim, 2009; Wu, & Chen, 2008, Arase, 2011). SMS spam has been overflowing in many countries, especially in India, Pakistan, and China (Delany, Buckley, & Greene, 2012). It is reported that the Chinese SMS clients received 16 spam messages every week in average, 17% were encountered SMS bombing in the second half of 2014(Center, 2014). In the USA, more than 69% of the surveyed mobile users claimed to have received text spam in 2012(Jiang, Jin, Skudlark, & Zhang, 2013). The content of spam generally is various kinds, such as advertisements, frauds, erotic services, etc. SMS spam serious violates customers’ rights and privacies, consumes the resource of mobile communication equipment, and result in the congestion of communication network (Sohn, Lee, Han, & Rim, 2012; Wadhawan & Negi, 2014; Wu et al., 2008, Martins, 2011).

At present, the technologies that are adopted by most mobile operators to filter SMS spam are transplanted from email spam filtering (Ma, Zhang, Wang, & Yu, 2016). However, different from email which consists of innate structure information such as subject, header, sources address etc., the SMS spam has the following special characteristics:

  • Limited length and fuzzy semantics: A SMS message is limited in 140 characters (Almeida et al., 2011), which results in a lack of context information (Sohn et al., 2012), and the spelling in SMS messages is usually at will by the users;

  • Large number of noisy terms: Except for normal words, the spam messages are usually injected in many kinds of abbreviations, symbols, cyber words, websites, variant words, and distort or deform sentences. Moreover, the spammers deliberately utilize error writing or spelling, and mispronunciation to avoid anti-spam system filtering. They also often use many kinds of symbol semantics (such as using ‘$’ or ‘¥’ to denote money) or word separators (such as ‘ca*sh’, ‘account#’, ‘tel 151#772#96’, etc.), These can cause difficulty for the word segmenting software;

  • Diversity: The content of SMS spams cover various domains, and refer to all sorts of themes. According to reports from the Chinese information security company-Qihoo 360, their users had carried on manual classifying for 6,130 million spam messages which were captured by their security software in 2014. It reflects that one of the largest category of SMS spam is sales advertising, reaching 69.1%, fraud and other miscellaneous illegal spam categories reached 10.2% and 7.6% respectively. In addition, the SMS spam relates to several ten fine-grained categories (360, 2014).

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 13: 2 Issues (2017)
Volume 12: 2 Issues (2016)
Volume 11: 2 Issues (2015)
Volume 10: 4 Issues (2014)
Volume 9: 4 Issues (2013)
Volume 8: 4 Issues (2012)
Volume 7: 4 Issues (2011)
Volume 6: 4 Issues (2010)
Volume 5: 4 Issues (2009)
Volume 4: 4 Issues (2008)
Volume 3: 4 Issues (2007)
Volume 2: 4 Issues (2006)
Volume 1: 4 Issues (2005)
View Complete Journal Contents Listing