Abstract:
Since most software systems are installed and configured by default, there are security
loopholes that could be exploited even though the majority of Internet users have not yet
developed a strong sense of security despite the fact that more data is being produced on
the Internet and on the connected nodes as the Internet continues to grow. Because to the
emergence of new methods for attacking networks, every host on the Internet may now be
considered a potential target. Because of this, the issue of the security of the information
sent across networks cannot be overlooked. The honeypot approach can not only be used
passively as an information system to cope with zero-day and future assaults, but it can
also be used actively to strengthen traditional defense systems against attacks in the future.
Honeypots may be used in both of these ways simultaneously. As a consequence of this,
we make use of machine learning in order to build a Smart Agent for Cyber Security Based
on Honeypot. In this, we will integrate a number of different machine learning approaches
with honeypot mechanisms so that we may anticipate the profile of the attacker and stop
them. In this section, we will compare the two different findings those achieved with and
those obtained without the assistance of honeypot and depict the results of the comparison
