...this.

In a sense, all data mining is phenomenal; it's just that the phenomenal part is usually done by hand. We want the computer to do the phenomenal part also.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...computer.

Even very young babies have a lot of innate knowledge of the world. My article The Well-Designed Child concerns what innate knowledge children probably do have about the world and what knowledge robots should be given. Elizabeth Spelke, [Spe94], investigates innate knowledge in babies experimentally.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...themselves.

It has been suggested that grouping baskets by customer is an example of clustering as treated in learning theory. This is incorrect, although there are some similarities. Consider two large identical baskets purchased ten minutes apart. Clustering would assign them to the same category, but these baskets would almost certainly have been purchased by different customers. Identical baskets purchased far enough apart would have an increased probability of having been purchased by the same customer, but it wouldn't be certain. Still, the literature on clustering might tell us something useful for the present problem.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.