|
Abstract:
|
With the advent of the modern computer, contemporary datasets can be too large to handle or analyze in traditional ways. An alternative approach is to summarize the data in such a way that ensures the dataset is more manageable yet also retains as much of the knowledge in the original dataset as possible. One consequence is that the dataset no longer consists of single point values in p-dimensional space but may be lists, intervals, distributions and so on, and are now hypercubes in space. Such data are called symbolic data. Some data are naturally symbolic in nature. We consider how such data arise, how they are structured, and how they can be analyzed. A comparison of symbolic analyses with standard analyses shows that the symbolic methods capture more of the inherent information contained within the symbolic data.
|
