Research On SMS Filtering Technology On Intelligent Mobilephone

In recent years, because of its advantage of accurate and reliable, rapid and timely,low price, SMS has developped rapidly.However, at the same time, it makes lots of spammessages to flood.This problem has been more and more serious, not only does it moreand more serious influences people’s normal life, even in a certain degree do harm to thesocial public security and stability.Therefore, the intercept and filter spam messages is a animportant problem to be resolved in today’s society. At present, however, SMS spamfiltering technology is mainly based on spam messages a certain characteristics of simplefilter, such as black and white list, keywords filtering and so on, which are being existed inobvious limitation and the defects of spam messages filtering efforts are not sufficient.In view of the current situation, it proposes a balanced based on the minimum riskNaive Bayes decision-making SMS spam filtering method. This method is based on thecontent of the text messages, by collecting a large number of normal text messages andspam messages examples, and combined with keywords technology, put them together asthe input of training of the Naive Bayes classification algorithm.And then will classifyingthe actual text messages.If the effect is not apparent, take through the typical features ofthe the spam messages to judge, and eventually got message category. And through theirown collection of text messages to makes a comprehensive experimental analysis to thismethod.The experimental results show that the method can accurately to SMSclassification, reduce the classification of legal messages error rate and improve the spammessages the recall ratio, and classification recall rate is above90%. Of course, thefiltering method still have various problems in the actual application,.We will improvethem in the future’s work to make it perfect.