Construction and management of automatical reasoning supported data mining metadata

With the development of data mining techniques, more and more data mining metadata are proposed for the sharing, exchanging, integration and standardization of data mining products. During the process of constructing data mining metadata, the evolution of data mining techniques, the different experiences and views of related of knowledge discovery techniques, the different experiences and data mining metadata lacks precise semantic due to their description with natural language and graphs, so the automatic consistency checking upon them has not been resolved well. In this research, Based on the novel formal logic DLR_DM in the description logic family, a data mining metadata with automatic reasoning AR-DMM is constructed in detail, and the corresponding algorithm is proposed. A formal reasoning framework is developed to automatically check the consistency of AR-DMM. A conflict detection example and a semantic consistency example are then illustrated. Results indicate that the AR-DMM algorithm is correct and effective on consistency checking on data mining metadata.