SMS Spam Interceptor
The efficient and accurate control of spam on mobile handsets is an important problem. Mobile spam incurs a cost on a per-message basis, degrades normal cellular service, and is a nuisance and breach of privacy. It is also a popular enabler of mobile fraud. In this project we implemented a novel method which incorporates the underlying byte-level data coding scheme of SMS to detect spam message on the access layer of mobile phone. We demonstrated using real-world spam messages that our scheme is robust, efficient and accurate to identify SMS spam. Based on our research we implemented our SMS Spam engine on Symbian powered mobile phone.
What we deliver
1. Automatic SMS spam Detection (unique solution)
2. SMS blocking via Keyword and number (existing solution in the market)
Whom we deliver
It is made as a final year project of university. and it is being used by friends. and it also won various nationwide events and earn huge applause internationally in Al-Zayed international IT competition Abu Dhabi 2011
Why is the project unique?
Most of existing Spam filtering techniques works on the application layer of mobile phone. Most of these techniques are straight forward adaptation of email spam detection schemes and usually incorporate features – specific words, character bi-grams and tri-grams for classification of spam messages. A well known shortcoming of these approaches is that their resource requirement (usually large memory and processing power) make them infeasible for deployment on resource constrained mobile phones. Moreover, these techniques can be easily evaded by word alteration, generating a local language SMS in roman English characters. Our Scheme operates on Byte-level Distributions using Hidden Markov models (HMMS) , which uses the underlying byte-level data coding scheme of SMS to detect spam messages. This scheme is robust to word adulteration techniques and language transformations as it works on the access layer of a mobile phone. The framework first builds a model of byte-level distributions of benign and spam messages and then builds benign and spam models using HMMs (Hidden Markov Models). This process leads to a new learning algorithm for the classification of spam SMS, which is based on the probabilistic variation from the trained models. Our framework is lightweight as it requires less processing and memory resources and hence can easily be deployed on mobile devices.
H.no 13 near Dar-Us-Salikeen Faizabad highway
Islamabad, Pakistan 44000