Download Exploratory Analysis for Efficient Data Mining

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Nonlinear dimensionality reduction wikipedia , lookup

Transcript
Exploratory Analysis for Efficient Data Mining
Thomas H. Burger, Eli Lilly and Company, Indianapolis, IN
Loren W. Burger, Jr., Mississippi State University, Starkville, MS
Abstract
Competitive advantage requires rapid formation of sound inferences for strategic information delivery. Key to knowledge formation is data
analysis and presentation. Exploratory analysis reveals relationships aiding selection of sites for data mining. Mining reveals unknown
relationships through methodological data selection, transformation, exploration and modeling. Further, it enables hypothesis testing and
recognition of patterns for decision support. SAS Institute (SI) advocates a framework (SEMMA) for exploration and mining which
systematizes the activities of sampling, exploring, modifying, modeling and assessing as an iterative stepwise process. Data exploration
and mining can be accomplished with code- or macro-based approaches. However, a tool-based approach delivered in a point-and-click
interface radically accelerates this process, speeding knowledge acquisition and inference formation. Previously restricted to those with
programming expertise, SAS Enterprise Minerï›› (EM) now enables non-programmers to mine. However, the need for reliable knowledge
and sound inferences dictates that statistical experts guide its use. We consider issues in tool use for exploration when conceptualizing a
data mining strategy.
-Paper will be provided during presentation-