Abstract– OnlineSocial Network (OSN) which plays a most important role in our day to day life,one user can communicate with one or more users by sharing several types ofinformation. Major issue in Online Social Network is to prevent user forposting unwanted messages such as vulgar, political words etc. There is need togive the users an ability to control the messages posted on their ownprivate/public space to avoid that unwanted contents to be displayed. Intoday’s available OSN System unwanted post will be directly posted on the userswall; to fill this gap, in this paper, I propose a system that automaticallyfilters the undesirable messages by allowing Online Social Network users tohave a direct control on the content posted on their walls. This is donethrough a filtering criteria to be applied to their walls, and a MachineLearning technique based soft classifier algorithm such as Radial Basis Function Network (RBFN).To do this, Black List (BL) mechanismis proposed in my system, which avoid undesired creators messages. BL is usedto determine which user should be inserted in Black List and decide when theretention of the user is finished. I have used DICOMFW as a special Facebookapplication due to which user can report a spam.
Keywords – Online Social Network (OSN), Machine Learning Techniques (MLT),Black List (BL),RadialBasis Function Network (RBFN), Content-Based Messages Filtering (CBMF), Short Text Classifier (STC)