Hybrid association-classification algorithm for anomaly extraction

Knowledge extraction is a process of filtering some informative knowledge from the database so that it can be used wide variety of applications and analysis. Due to this highly efficient algorithm is required for data mining and for accessing data from large datasets. Although there are various techniques implemented for the detection of anomalies using frequent item sets using apriori algorithm but the technique applied are not suitable for large database and contains more error rate and also the classification ratio is less. Hence in this paper an efficient technique is implemented using the combinatorial method of Classification and association rule mining. First the fuzzy apriori algorithm is applied to generate frequent item sets and then CART algorithm is applied for the classification of the network anomalies.