The outlier-based method provides a promising solution for the criminal incident association problem, and it could be generalized to other application areas. When applying this method to datasets with more records and more attributes, some adjustments are required to keep this approach efficient and still effective. For example, some powerful commercial OLAP software programs can be introduced to speed up generating the cells. Additionally, some faster attribute selection/reduction methods should be considered. The outlier-based association method itself may also need some adjustments when dealing with large datasets, such as a top-down cell building strategy or some early termination rule. With these improvements the method could have even broader applicability.