Data Mining

DSA's Approach To Data Mining

DSA’s take on Data Mining is that the primary goal is to provide actionable business intelligence by presenting information on significant trends, patterns and relationships in company operations as well as the marketplace. To be effective, the Data Mining process must be flexible to meet changing needs, and timely to support ongoing decision making. Most practitioners agree that Data Mining can be divided into two distinct applications.

Knowledge Discovery

One application is “Knowledge Discovery” which involves developing descriptive information which summarizes relevant results. Knowledge Discovery can be implemented via a variety of techniques such as standardized reporting, ad-hoc queries and reporting, drill-down techniques, descriptive statistical techniques, exploratory data analysis, decision trees, etc.

Forecasting

The other application is “Forecasting” (or predictive modeling) which provides predictions of the probability of future events such as responding to a promotion, developing into a high value customer, belonging to a given customer base sub-group, etc. DSA supports both types of Data Mining applications and has significant experience in implementing “real world” applications.

Data Mining Tools

While we are constantly evaluating new software and analytical architectures, our basic set of analytical tools includes:

SAS Ver 9 (Base, STAT, ETS, QC, OR, etc.)
KnowledgeSeeker (CHAID Decision Trees & Ad-Hoc Reports)
Answer Tree (CHAID Decision Trees & Ad-Hoc Reports)
KnowledgeStudio (Advanced Modeling Tools)
A number of proprietary software tools
Alterian DBMS with advanced ad-hoc query and analytical reporting capabilities
SAS end user oriented tools (Enterprise Guide, Text miner, OLAP, etc.)