الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 84 |  |  |
| | | from | 84 | | |
Abstract
No one can deny that the online social media becomes a main part from
our life; it has become indispensable. Nowadays, it is not used only for
communication or entertainments however it plays a great role in a lot of fields
such as education, trading, advertising, politics, media, and economy. It drew
the attention of all people. It has increased in the number of users. So, the
security of the oniine Social Networks (OSN) gains a great importance from
day to day. So, we start to compare between the two most important sites on
OSN: Facebook and Twitter in order to decide which one was be affected by the
spam and malicious accounts much than the other one. We find that spam and
malicious accounts have great dangerous on Twitter than Facebook. That is
because the main target of Twitter is to get the real-time news and trends topics
that may easily affected by spam and malicious accounts. On the other hand,
the main target of Facebook is to communicate and make relations. Both sites
may be affected by the abnormal accounts but their danger is affected on
Twitter is more than Facebook.
This thesis does not open new field instead it will change the used
strategies to solve the existence problem. Most of researchers try to solve the
problem of spam and malicious accounts on the OSN. Most of the presented
solutions lack for the flexibility to deal with the new features and categorize
them to normal or abnormal features. To be able to provide new approach,
there are several stages. First, the thesis starts to make a comprehensive study
on the features of the normal and abnormal accounts, to be able to differentiate
between them. The thesis categorizes the abnormal features into six groups:
BOT behavior, the observation of URL Features, Timeline content features, the
profile properties, Cheating features, and Analysis of the activities features. By
the observation and some facts, the thesis can find eleven features which called
them crucial Normal/Abnormal features. If one of these features in account is
founded in account, the approach can directly categorize it without need to use
complicated process. Using the crucial features will help a lot in reducing the
.