SpamNet – Spam Detection Using PCA and Neural Networks

This paper describes SpamNet – a spam detection program, which uses a combination of heuristic rules and mail content analysis to detect and filter out even the most cleverly written spam mails from the user’s mail box, using a feed-forward neural network. SpamNet is able to adapt itself to changing...

Full description

Saved in:
Bibliographic Details
Published inIntelligent Information Technology pp. 205 - 213
Main Author Lad, Abhimanyu
Format Book Chapter Conference Proceeding
LanguageEnglish
Published Berlin, Heidelberg Springer Berlin Heidelberg 01.01.2004
Springer
SeriesLecture Notes in Computer Science
Subjects
Applied sciences
Artificial intelligence
Bayesian Classification
Computer science; control theory; systems
Exact sciences and technology
Extractor Module
Feature Vector
Learning and adaptive systems
Mailing List
Neural Network
Energy analysis
Filter
Classification
Heuristic method
Cryptanalysis
Electronic mail
Neural network
Feedforward
Artificial intelligence
User behavior
Principal component analysis
Content analysis
Online AccessGet full text

Cover

More Information
Summary:This paper describes SpamNet – a spam detection program, which uses a combination of heuristic rules and mail content analysis to detect and filter out even the most cleverly written spam mails from the user’s mail box, using a feed-forward neural network. SpamNet is able to adapt itself to changing mail patterns of the user. We demonstrate the power of Principal Component Analysis to improve the performance and efficiency of the spam detection process, and compare it with directly using words as features for classification. Emphasis is laid on the effect of domain specific preprocessing on the error rates of the classifier.
ISBN:9783540241263
3540241264
ISSN:0302-9743
1611-3349
DOI:10.1007/978-3-540-30561-3_22