Disorder measures supporting autonomous, imprecise database mining

There is a basic problem of database analysis: there are huge amounts of data, more than can ever be analyzed by humans. The question is how to computationally identify and present the most useful information. Discovering the details of correctly anticipated information is important. However, perhaps more important, is recognizing the unexpected. While the greatest possible return might result from serendipitously browsing all possible sources, this is infeasible due to resource constraints. The discussed approach sifts out non useful data until useful information is discovered. This approach is in contrast to existing approaches assume that they can select and rank order the most useful data when confronted by all the data, useful and non useful. The problem with this is that it is computationally infeasible to look at all the data; so, heuristic choices are made that strongly constrain what might be discovered. The heuristic assumption is that reducing cognitive dissonance increases useful information. The speculation is that database exploration can be accomplished through a progressive reduction of cognitive dissonance. This will be done by progressively discarding attributes that have limited information value and by partitioning the data to increase information within the resulting partitions