Knowledge Discovery Method by Gradual Increase of Target Baskets from Sparse Dataset

Summary

Data mining is one of the methods to extract some knowledge from large amount of data and KeyGraph is one of the unique methods for data mining. The result of KeyGraph analysis is shown like network diagrams; the analyst tries to understand the meaning of links and make reasonable scenarios. In this process, the more complex the link structure is, the more difficult to understand its meaning. For this difficulty, we developed a preprocessing method enabling to generate simpler link structure at first and also generate more complex structure gradually. By this method, we could obtain more detailed and various scenarios.