Hybridized Data Technique for Evaluate the Anomaly

摘要

Intrusion behavior and detection analysis particularly rely upon the type of data. Most of the datasets used in intrusion analysis are heterogeneous and imbalanced data sets. In these data sets, the features vary with a huge difference in between and within the feature values. This is very effective while taking decision, especially in the supervised learning. To analyze the intrusion problem, support vector machine uses distance calculation. But data set contains both numerical and nominal values. Here, we have replaced these nominal data points with the appropriate values and balanced the dataset. This method enhances the detection rate of the anomaly detection system.