Advances in Spam Filtering Techniques

摘要: Nowadays e-mail spam is not a novelty, but it still an important rising problem with big economic impact in society. Fortunately, there are different approaches able to automatically detect and remove most of those messages, the best-known ones based on machine learning techniques, such as Naive Bayes classifiers Support Vector Machines. However, several models filters, something literature does always acknowledge. In this chapter, we present compare seven versions classifiers, well-known linear Machine new method Minimum Description Length principle. Furthermore, have conducted empirical experiment six public real non-encoded datasets. The results indicate that proposed filter easy implement, incrementally updateable clearly outperforms state-of-the-art filters.