22-05-2012, 10:41 AM
A NEURAL NETWORK APPROACH TO DETECT SPAM
ABSTRACT:
Email occupies server storage space and consumes network
bandwidth in order to overcome this problem email readers spend a significant amount of
time to detect and deleting spam (Junk E-mail) messages. There is a need to develop
solutions that are automatically distinguishing legitimate E-mail from spam. One
approach that has been used so far to detect spam is Naïve Bayesian approach. This
approach consists of large feature sets of binary attributes that determine the existence of
common keywords in spam. However spam applications have also recognized these
attempts and have develop tactics tool circumvent to overcome this filters. In this paper a
Neural Network approach for detecting spam is described. This approach uses for feature
set containing descriptive characteristics of words and messages similar to those that a
human reader would use to identify spam. This Neural Network approach uses fewer
features set than a Bayesian approach.
K.SELVANI DEEPTHI
M.TECH(CSE),III SEMESTER
ANDHRA