|
|
|
|
|
|
 |
|
|
Crowd-Sourcing Insight from Digital Entities
|
|
|
|
|
|
| Organization: | Evolved Analytics, LLC |
|
|
|
|
|
|
Wolfram Technology Conference 2012
|
|
|
|
|
|
Champaign, Illinois, USA
|
|
|
|
|
|
The evolutionary search and the core of DataModeler rewards models which are both simple and good approximations of the supplied input-response data. This selection pressure will automatically focus on key factors (variables) and produce relatively concise global models and handle real-world situations wherein the inputs are coupled or there are many variables and relatively few observations.
Because there are an infinite number of models which can (approximately) capture the behavior of a finite data set, diverse model forms will be synthesized. We value this diversity because it forms the foundation of trustable models which can detect extrapolation and changes in the targeted system. We also value this diversity because we can survey the best and brightest and look for combinations and couplings between inputs that are ingredients for success.
In summary, diversity leads to insight as well as effective data models. In this talk, we will illustrate that for real-world applications as well as the advantages relative to the single-minded focus embraced by conventional data modeling techniques.
|
|
|
|
|
|
|
|
|
|
|
|
http://www.wolfram.com/events/technology-conference/2012
|
|
|
|
|
|
|
| | | | | |
|
For the newest resources, visit Wolfram Repositories and Archives »
