A System to Filter Unwanted Messages from OSN User Walls

ABSTRACT One fundamental issue in today On-line Social Networks (OSNs) is to give users the ability to control the messages posted on their own private space to avoid that unwanted content is displayed. Up to now OSNs provide little support to this re uirement. !o fill the gap" in this paper" we propose a system allowing OSN users to have a direct control on the messages posted on their walls. !his is achieved through a fle#ible rule-based system" that allows users to customi$e the filtering criteria to be applied to their walls" and a %achine &earning based soft classifier automatically labeling messages in support of content-based filtering. 'nde# !erms( On-line Social Networks" 'nformation )iltering" Short !e#t *lassification" +olicy-based +ersonali$ation. Existing System ,e believe that this is a key OSN service that has not been provided so far. 'ndeed" today OSNs provide very little support to prevent unwanted messages on user walls. )or e#ample" )ace book allows users to state who is allowed to insert messages in their walls (i.e." friends" friends of friends" or defined groups of friends). -owever" no content-based preferences are supported and therefore it is not possible to prevent undesired messages" such as political or vulgar ones" no matter of the user who posts them. +roviding this service is not only a matter of using previously defined web content mining techni ues for a different application" rather it re uires to design adhoc classification strategies. !his is because wall messages are *onstituted by short te#t for which traditional classification %ethods have serious limitations since short te#ts do not +rovide sufficient word occurrences. Proposed System !he aim of the present work is therefore to propose and e#perimentally evaluate an automated system" called )iltered ,all (),)" able to filter unwanted messages from OSN user walls. ,e e#ploit %achine &earning (%&) te#t categori$ation techni ues ./0 to automatically assign with each short te#t message a set of categories based on its content. !he ma1or efforts in building a

Contact: 040-40274843, 9703109334 Email id: academicliveprojects@gmail.com, www.logicsystems.org.in

A System to Filter Unwanted Messages from OSN User Walls

robust short te#t classifier are concentrated in the e#traction and selection of a set of characteri$ing and discriminate features. !he solutions investigated in this paper are an e#tension of those adopted in a previous work by us .20 from which we inherit the learning model and the elicitation procedure for generating pre-classified data. !he original set of features" derived from endogenous properties of short te#ts" is enlarged here including e#ogenous knowledge related to the conte#t from which the messages originate. 3s far as the learning model is concerned" we confirm in the current paper the use of neural learning which is today recogni$ed as one of the most efficient solutions in te#t classification ./0. 'n particular" we base the overall short te#t classification strategy on 4adial 5asis )unction Networks (45)N) for their proven capabilities in acting as soft classifiers" in managing noisy data and intrinsically vague classes. %oreover" the speed 6 in performing the learning phase creates the premise for an ade uate use in OSN domains" as well as facilitates the e#perimental evaluation tasks.

IMPLEMENTATION 'mplementation is the stage of the pro1ect when the theoretical design is turned out into a working system. !hus it can be considered to be the most critical stage in achieving a successful new system and in giving the user" confidence that the new system will work and be effective. !he implementation stage involves careful planning" investigation of the e#isting system and it7s constraints on implementation" designing of methods to achieve changeover and evaluation of changeover methods.

Contact: 040-40274843, 9703109334 Email id: academicliveprojects@gmail.com, www.logicsystems.org.in

A System to Filter Unwanted Messages from OSN User Walls

Modules: 1 !iltering rules

'n defining the language for )4s specification" we consider three main issues that" in our opinion" should affect a message filtering decision. )irst of all" in OSNs like in everyday life" the same message may have different meanings and relevance based on who writes it. 3s a conse uence" )4s should allow users to state constraints on message creators. *reators on which a )4 applies can be selected on the basis of several different criteria8 one of the most relevant is by imposing conditions on their profile7s attributes. 'n such a way it is" for instance" possible to define rules applying only to young creators or to creators with a given religious9political view. :iven the social network scenario" creators may also be identified by e#ploiting information on their social graph. !his implies to state conditions on type" depth and trust values of the relationship(s) creators should be involved in order to apply them the specified rules. 3ll these options are formali$ed by the notion of creator specification" defined as follows. " Online setup #ssist#nt $or !Rs t%res%olds:

3s mentioned in the previous section" we address the problem of setting thresholds to filter rules" by conceiving and implementing within )," an Online Setup 3ssistant (OS3) procedure. OS3 presents the user with a set of messages selected from the dataset discussed in Section ;'3. )or each message" the user tells the system the decision to accept or re1ect the message. !he collection and processing of user decisions on an ade uate set of messages distributed over all the classes allows to compute customi$ed thresholds representing the user attitude in accepting or re1ecting certain contents. Such messages are selected according to the following process. 3 certain amount of non neutral messages taken from a fraction of the dataset and not belonging to the training9test sets" are classified by the %& in order to have" for each message" the second level class membership values.

Contact: 040-40274843, 9703109334 Email id: academicliveprojects@gmail.com, www.logicsystems.org.in

A System to Filter Unwanted Messages from OSN User Walls

&

Bl#'(lists:

3 further component of our system is a 5& mechanism to avoid messages from undesired creators" independent from their contents. 5&s are directly managed by the system" which should be able to determine who are the users to be inserted in the 5& and decide when users retention in the 5& is finished. !o enhance fle#ibility" such information are given to the system through a set of rules" hereafter called 5& rules. Such rules are not defined by the SN%" therefore they are not meant as general high level directives to be applied to the whole community. 4ather" we decide to let the users themselves" i.e." the wall7s owners to specify 5& rules regulating who has to be banned from their walls and for how long. !herefore" a user might be banned from a wall" by" at the same time" being able to post in other walls. Similar to )4s" our 5& rules make the wall owner able to identify users to be blocked according to their profiles as well as their relationships in the OSN. !herefore" by means of a 5& rule" wall owners are for e#ample able to ban from their walls users they do not directly know (i.e." with which they have only indirect relationships)" or users that are friend of a given person as they may have a bad opinion of this person. !his banning can be adopted for an undetermined time period or for a specific time window. %oreover" banning criteria may also take into account users7 behavior in the OSN. %ore precisely" among possible information denoting users7 bad behavior we have focused on two main measures. !he first is related to the principle that if within a given time interval a user has been inserted into a 5& for several times" say greater than a given threshold" he9she might deserve to stay in the 5& for another while" as his9her behavior is not improved. !his principle works for those users that have been already inserted in the considered 5& at least one time. 'n contrast" to catch new bad behaviors" we use the 4elative )re uency (4)) that let the system be able to detect those users whose messages continue to fail the )4s. !he two measures can be computed either locally" that is" by considering only the messages and9or the 5& of the user specifying the 5& rule or globally" that is" by considering all OSN users walls and9or 5&s.

Contact: 040-40274843, 9703109334 Email id: academicliveprojects@gmail.com, www.logicsystems.org.in

A System to Filter Unwanted Messages from OSN User Walls

System Re)uirement Spe'i$i'#tion: *#rd+#re Re)uirements: +rocessor Speed 43% -ard ?isk )loppy ?rive Aey 5oard %ouse %onitor So$t+#re Re)uirements: Operating System 3pplication Server )ront Fnd Scripts Server side Script ?atabase ?atabase *onnectivity B ,indowsC29CD96@@@9E+ B !omcat2.@9>.E B -!%&" Gava" Gsp B GavaScript. B Gava Server +ages. B %ys l B G?5*. =.= :-$ 62> %5 (min) +entium <'''

- 6@ :5 =.// %5 Standard ,indows Aeyboard

- !wo or !hree 5utton %ouse S;:3

Contact: 040-40274843, 9703109334 Email id: academicliveprojects@gmail.com, www.logicsystems.org.in