| Summary: | 碩士 === 雲林科技大學 === 資訊管理系碩士班 === 96 === Purpose - Spam has been a normal phenomenon, and it will be inconvenient if lots of unsolicited bulk SpaSMS messages are sent to user’s mobile phone as spam in email. The investigation of SpaSMS is rare, and therefore we try to propose a novel personalized and artificial immune system added SpaSMS filtering system.
Design/methodology/approach - This study tries to use two classifiers to filter SpaSMS. One is generalized classifier, and the other is personalized classifier. This research uses Naïve Bayesian algorithm to build generalized classifier and utilize artificial immune system to extract SpaSMS’s features. Personalized classifier is also created for individual through k-Nearest Neighbor algorithm.
Findings - Artificial immune system and personalized filtering can improve the performance of SpaSMS filtering.
Originality/value - This research takes the leading in using a total system solution in SpaSMS filtering. Moreover, this study is the first to use artificial immune system in SpaSMS’s feature extraction. In addition, this study is the first to combine generalized classifier with personalized classifier to filter SpaSMS. Thus, it can provide a better performance than Naïve Bayesian and decision tree algorithms.
|
|---|
