DATABASE SECURITY INCURSION RECOGNITION TECHNIQUE USING NEURAL NETWORK

Abstract

Database Intrusion Detection System (IDS) is an expert system looking for evidence of attacks on known vulnerabilities of the system. It holds a statistical model of the behaviour of a user on a system under surveillance. There are several techniques, protocols, and algorithms to increase the security level of database. In such works, there is a lack of time complexity analysis of the techniques. This time complexity has occurred due to the comparison process carried out at each time the user query is given i.e., comparing the profiles of online transactions and the stored authorized transactions each time when the query is received. This system time complexity also affects the system performance in terms of their precise security. It learns the habits a user working with the computer and to raise warnings when the current behaviour is not consistent with the previous learnt patterns, thus detecting whether the user is authentic or not. The system can be implemented using MATLAB. MATLAB is a numerical computing environment. It allows matrix manipulations, plotting of functions and data implementation of data. The learning process by neural network avoids the unauthorized transactions in the DBMS and reduces the time complexity the project improves the performance of the database system. The neural network implementation will show the effectiveness of the proposed IDS technique in securing the database from the intruders. The performance of the proposed technique is evaluated by utilizing different statistical performance measures.