基于SVM的电子邮件分类系统研究-计算机软件与理论专业论文.docxVIP

山东师范大学硕士学位论文速度，还要对特征词筛选方法进行系统研究，挑选出最适合邮件过滤的特征项选择方法， 山东师范大学硕士学位论文 速度，还要对特征词筛选方法进行系统研究，挑选出最适合邮件过滤的特征项选择方法， 为提高分类效果需要广泛收集邮件样本进行训练。论文肯定还有许多不完善的地方，相 关工作还有待进一步研究。 关键词二电子郎件；支持向量机；分类；特征选择 分类号：TP391，1 山东师范大学硕士学位论文The 山东师范大学硕士学位论文 The Research of Email Classification Based on SVM ABSTRACT E-mail is the most widely used Intemet application，the most popular network functions． With the popularity‘of information technology．It has now evolved into a much more complex and rich system allow the transmission of voice，pictures，i。mages，multimedia files and other information．Report on the accounts or even if the database can be circulated on the Intemet in theform of e．mail attachments．Now,e-mail has become the life blood of many businesses and organizations．Users carl manage project through e—mail discussions．Sometimes under the fast or the exchange of intercontinental e-mail messages we make our decisions．However,wi也all increase in the number of e-mail，how t‘o effective classify e-mail and filter out spam bothers many people． Support Vector Machine is a new generation of learning machines based on statistical learning theory．It has many attractive features and its ability to：function,learning ability and efl]ciency mnst be superior 10 the traditional artificial neural networks．The past 1 0 years， Vapnik and his colleagues put。forward SVM algorithm based on statistical learning theory．In small samples，nonlinear and high-dimensional pattern recognition it has some unique advantages．The method can also be applied to other maeh．ine learning problems．Many scholars believe that it is becoming hot new field of machine learning after pattem recognition and neural network．SVM will promote a significant development on machine learning theory and technology．SLT and SVM made encouraging progress 011 application of the kernel in dewing 、Ⅳi血the small sample problem．It has been the best learning theory on the small sample statistical estimates and forecast study． This essay research on email classif