Effective Intrusion Detection Using Feature Selection : A Machine Learning based Approach

Internet security is a serious concern in the present era. There is an enormous growth of the Internet in terms of its usage and resources. Almost all major organizations, educational institutes, governments and individuals are dependent upon the Internet for providing their services. Therefore, smooth running of the Internet and maintaining integrity & confidentiality of the sensitive information over the Internet is the most important aspect of the growth of the information based organizations. However, presence of flaws in the most popular software provides numerous chances for malicious users to mount a variety of attacks to disrupt services and integrity of sensitive information over the Internet. Thus, the security of the Internet resources has become a very serious issue and a global concern for the governments, commercial organizations and individuals. Moreover, due to the advancement of computers, the behavior of intrusions has become more complex. Machine learning techniques have evolved as a promising solution. However, performance is affected due presence of a large number of redundant and irrelevant features. Feature selection helps to identify most relevant features.