Network Intrusion detection and Fake information detection system using Bayesian estimation
Hundreds of billions are spent every year around the world to secure network systems and prevent attacks. Intrusion detection has recently become a crucial technology for security mechanisms. Making the difference between false and true information on social media is a difficult undertaking given the vast amount of data available. In reality, multinomial distributions are used in several statistical models that deal with this issue. present a novel approach to neural networks in an intrusion detection system (IDS). It entails creating a model of reference approach and an evaluating difference between current and standard behavior using a classification technique of Bayesian linked to a learning unsupervised algorithm. Instead of minimizing the likelihood of error, the defender's aim is to reduce the average system estimation. The cost functions were derived from the suggestions of the estimate mean square error (MSE) matrices. The proposed detection estimation design improves classic sensors include the traditional detector of Bayesian and in terms of overall MSE, a chi-squared analyzer makes a substantial difference, according to numerical results. This paper aims to present an Intrusion detection is discussed, and also a new method for intrusion detection method based on an adaptable Bayesian algorithm. With excellent detection accuracy and a rapid reaction time, this method properly classifies different types of attacks in the KDD99 is a benchmark sample for intrusion detection. The experimental results further show that for intrusion detection, this approach maximizes the detection rate (DR) while minimizing the false positive rate (FPR).