Combined Mining Approach to Generate Informative Patterns

Abstract Business data mining applications involve huge amount of heterogeneous, distributed data. In such a case to use traditional data mining algorithms for obtaining comprehensive information about business, which will be helpful for decision making, is very time and space consuming. Traditional data mining methods involve single step data mining process to generate patterns and also they deal with homogeneous features of dataset. They need to follow join operation to get useful information from multiple large data sources. We consider Combined Mining as an approach to generate more informative patterns by considering multiple data sources or multiple features or multiple methods. Here we are going to discuss multifeature combined mining and multimethod combined mining methods. In multifeature combined mining, we obtained pair patterns, incremental pair patterns and cluster patterns by considering multiple heterogeneous features from data sources. In multimethod combined mining approach, multiple data mining methods has been used to generate more informative knowledge.