Dimensionality reduction in text classification for spam detection using recursive feature elimination
Nowadays many devices having the capability of receiving emails worldwide. Since there is a large number of the availability of accessing email, this also affected to the distribution of spam and other malicious threats that primarily use email as its carrier. Therefore the intensity of countermeasures for this kind of threat must be increased. Web mining is one technique to collect data from the internet. Spatio-temporal data of email messages can be collected using web mining so the useful information in order to detect a spam within email messages can be extracted for further analysis. This study aims to evaluate the possibility of an alternative solution to classify an email whether it is a spam or a non-spam category, by using recursive feature elimination (rfe) together with support vector machine (svm). The result shows that svm- rfe is able to classify spam and non-spam messages in 100% accuracy using 11 features.
M00179 | (wh) | Available |
No other version available