ABSTRACT
Online Social Network (OSN) is a vast environment where people communicate with each other and share their ideas. The ideas that are shared by the person is posted on the user profile. So that all the views of the person are shared among the people. The major drawback in OSN is the irrelevant messages are shared in the user’s wall without the knowledge of the user. The user doesn’t require these irrelevant messages on their wall, so it need to be filtered. This problem can be overcome by creating an Ontology that includes set of rules based on the user’s interest to avoid irrelevant messages. Ontology is a generic knowledge that represents agreed domain semantics that can be reused by different kinds of applications or tasks. The content that are posted in the wall are analyzed using content analysis method. This paper proposes a method for the online social network using ontology to clear unwanted messages posted in the wall.
Index Terms: Ontology, Content analysis, Rule based system.
INTRODUCTION
DATA MINING
Data mining is the process to extract the information from the set of data and transform it into a user understandable form. Data mining means a knowledge discovery in database. It is the process of analyzing data and change it into a user readable form. The process of data mining include selection of target data from the data set, then the data are processed and is transformed into understandable form by transformation and it is mining into patterns and finally the information are evaluated and stored in the knowledge base. The techniques used in the data mining are Estimation, Prediction, Classification, Clustering and association rules.
SEMANTIC WEB
Semantic web is not a separate web, it is the extension of already existing one, which provide information in a well defined form. Semantic web is used to convert the unstructured or semi structured data into a web of data. Semantic web content are easilt accessed by the humans, but it is not easily accessed by the computers. Semantic web are build using the similar meaning, patterns, structures, relations from the existing web. Semantic web are used in data mining with the help of ontology. Ontology is the backbone of semantic web, by using ontology the semantic web are transformed into understandable form. Some of the SW technologies are XML and RDF.
ONTOLOGY
Ontology is study of nature and categories of being and their relationships. Ontology act as backbone to semantic web. The web content are not understand by the user, to transform the web content into user understandable form ontology is used in data mining. Ontology is a technique that is used in the semantic web to do the transformation. Ontology is not only used to semantic web transformation but also for defining the classes, relations and properties of each individual.
CURRENT RESEARCH
A wide research is being done in the Ontology technique.
Gene Ontology
The Gene Ontology project provides an ontology defined terms representing gene product properties. The Gene Ontology is structured as a directed acyclic graph and each term has defined relationships to one or more other terms in the same domain and to other domains. The Ontology covers three domains: Cellular Component, the parts of a cell or its extracellular environment. Molecular Function: The elemental activities of a gene product at the molecular level, such as binding. Biological Process: Operations or sets of molecular events with a defined beginning and end to the functioning of integrated living units.
Gene Ontology (GO) is a controlled biological terminology being created by a consortium of bio-infomaticians. Even though it’s relatively new to the world when compared to other ontologies, GO has greater impact on bioinformatics community. GO started with terminologies from three genomic databases: Flybase, Saccharomyces Genome Database and Mouse Genome Database and has developed three hierarchies of terms to describe biological processes, cellular components and molecular functions. Gene definitions and comments by the authors are given as annotations in the ontology.
Existing system
The existing system is a method in Online Social Networks (OSNs) is to give users the ability to control directly on the unwanted messages posted on their own profile. This is carried out a rule based system that controls the filtering criteria applied to the user wall. The automatic labeling of messages is done using Machine Learning based soft classifier and content based filtering. The aim of the present work is therefore to propose and experimentally evaluate an automated system, called Filtered Wall (FW), able to filter unwanted messages from OSN user walls.
Online social networking are gaining more and more attention since they provide a platform to explore one’s own ideas. However the major disadvantage in OSN is the user wall is filled with the content that out of user’s interest.
The content that are posted in the user wall or the Graphical user interface are collected and given for the filtration process. This is one of major issue in the OSN, which makes the user to distract from their view and makes the user to avoid using OSN’s. This method id useful to avoid those irrelevant messages out of user interest.
The function of this method is to avoid those unwanted messages and to provide the content that are needed by the user or based on the user interest. After the content enters into the user wall it is sent through the Machine learning based soft classifier which is used to extract the metadata from the content of the message. Filtering wall uses the metadata that are provided by the machine learning classifier and also the data extracted from the user’s wall and the social groups is given to the Content based Filtering and with the support of Blacklist rules. Based on the result of the content based filtering the message will be published in the user wall or again it is used by the filtering wall. This method often leads to more time consuming and a complex process to achieve it.
Content Based Filtering
Content filtering systems are designed to classify a stream of dynamically generated information that are given by the information producer and present to the user those information that are likely to satisfy their requirements. In content-based filtering, each user is assumed to operate independently. As a result, a content-based filtering system selects information items based on the correlation between the content of the items and the user preferences.
Machine Learning-Based Classification
The first-level classifier performs categorization of the labeled messages. The classifier performs partitioning of labeled messages assigning those message to each of the labeled classes. Among the variety of multiclass ML models the RBFN model is chose.
Filtering rules:
For FRs specification, the three main issues that are consider. In OSNs the same message may have different synonyms and significance based on who writes it. FRs are used to create constraints on the message posted based on the user interest. One of the way to create FR is based on one’s profile attributes. The user’s can also be identified by exploiting information on their social graph. This implies to state conditions on the relationships between the creators and the user. The FR is the action that the system has to perform on the messages that satisfy the rule. The possible actions that are considered are ‘block’ and ‘notify’.
Blacklists:
Blacklist is a mechanism used to avoid the messages that are created by the undesired creators independent of the content. BL are used to determine which users are to be inserted and what content are to be inserted. BL is managed directly by the system. BL rules are formed by the user based on their opinion to decide who has to be banned and for how long from their wall. Similar to FRs, our BL rules make the wall owner able to identify users to be blocked according to their profiles as well as their relationships in the OSN. Therefore, by using the BL rules, the wall owners are able to ban the creators based on their relationships or based on their bad behaviors. This can be done for a specified time or for an undetermined time. Moreover the rules specifies that the banned users have to stay in the BL for the specified time period, the user can also enlarges the time period to ban the creator from the user in the OSN.
PROPOSED WORK
Online Social Networks (OSN) is one of the important communication media to share several types of content, including free text, image, audio, and video data. There is a possibility of posting or commenting other posts on particular public/private areas, called in general as walls. Information filtering can therefore be used to give users the ability to automatically control the messages posted on their own walls, by filtering out unwanted messages. Rules are specified using Ontology to filter the messages that are blocked by the users. The users may have interest on specific domain that will be added into the account of the users. Message posted by other users will be of any domain, but the system will post the message to the user’s wall that belongs to the user’s interest. Messages that are not related to user’s interest will be blocked and will not be posted in the wall.
ARCHITECTURE
CONCLUSION:
Ontology is created based on the domain knowledge. An ontology is a catalogue of the types of things that are assumed to exist in a domain of interest D from the perspective of a person who uses a language L for the purpose of talking about domain. Ontologies are used not only to represent a domain of interest, but also define concepts, describe relations among them and insert individuals. The existing system uses a method for filtering unwanted messages in a social networking using machine learning. To filter the unwanted messages posted on the user wall in a social networking content based filtering method is used. The proposed system uses ontology concept to filter the messages. It is done by creating the rules in order to block the unauthorized user and unwanted messages posted by the people on the user wall. The messages posted on the wall is analyzed by the content analysis method. Then the content is filtered based on the rules that are specified by the user. This provides a better result than the machine learning technique. Unwanted messages, unauthorized people and also the advertisements can also be blocked in the future.
REFERENCES:
1. http://en.wikipedia.org/wiki/Data_mining
2..http://www.findwhitepapers.com/
technolgy/data_management /data_mining
3.http://en.wikipedia.org/wiki/ Semantic_Web
4. Sankaranarayanan, J., Samet, H., Teitler, B. E., Lieberman, and M. D., Sperling, J. TwitterStand: news in tweets. In Proc. ACM GIS’09 (Seattle, Washington, Nov. 2009), 42-51.
5. K. Strater and H. Richter, ‘Examining Privacy and Disclosure in a Social Networking Community,’ Proc. Third
Symp. Usable Privacy and Security (SOUPS ’07), pp. 157-158, 2007
6. Sankaranarayanan, J., Samet, H., Teitler, B. E., Lieberman, and M. D., Sperling, J. TwitterStand: news in tweets. In Proc. ACM GIS’09 (Seattle, Washington, Nov. 2009), 42-51
...(download the rest of the essay above)