Haiku is currently applied to fixed size tuples of mixed nominal and numeric data. However, the data-data clustering techniques used can be applied to any data type where a measure of similarity between two items can be calculated. This is simply a function which takes two data items and returns a numerical measure which corresponds with the intuitionistic concept of similarity between those two items. The current similarity metric is domain independent, but domain knowledge could be used to improve the calculation of similarity.
Only classification rules are discussed in this paper, but Haiku is easily extended to the visualisation of association rules. Indeed, any form of knowledge can be visualised, providing that a numeric measure indicating the degree of agreement between that knowledge and an item of data can be calculated.