Basket Analysis on Meningitis Data

Basket Analysis is the most representative approach in recent study of data mining. However, it cannot be directly applied to the data including numeric data. In this paper, we claim the importance of the selection and the discretization of numeric attributes in the data preprocessing stage for the wider application of Basket Analysis, and propose the algorithm and the performance measures for the selection and the discretization. Through the application of the developed method and Basket Analysis to the meningitis data, their performance is evaluated, and the desirable feature of the performance measure is clarified.