ABSTRACT
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
There are a myriad number of data mining tasks that can be performed on sets of vectors with continuous vector components, i.e., whose values are real numbers. For wide vectors corresponding to high-dimensional spaces, it might be useful to filter out individual vector components that do not contribute to the data patterns being studied to make the problem easier to work with and to reduce noise (see Chapter 10 on dimension reduction). Perhaps the vectors are each assigned a discrete property, and we want to predict properties of new, unseen vectors (see Chapter 9 on classification).
