News | Organizations | Economic | Documentation | Newscenters | Miscellaneous | Espionage

Intelligence organizations

MINING FOR INTELLIGENCE

Data mining tools are intelligent pieces of software which use a variety of complicated techniques to reveal fundamental patterns in information. They've been described as "virtual data detectives". These technologies penetrate the surface of day-to-day data and go deep into information bases and warehouses to discover patterns, connections, trends and relationship - some obvious and some not - and then present the finding in a meaningful, often graphic, way.

IBM has launched "net.Mininf" services to help companies chart and predict the buying behavior of customers.

Data mining is still in the pilot phase, and is aimed only at discreet information systems or subsets of marketing and customer databases.

Using data mining techniques to show hidden associations, predictable sequences and develop accurate classifications of customers, suppliers, or products, companies can increase the business value of their information assets manifold.

Methods for data mining:

• Cluster analysis: identifies associations between different "data points";

• Linkage analysis: used to link various data points or events together (can show when separate product purchases are related;

• Time series analysis (sequencing): links various events in time (financial annalists use the method to predict interest rate fluctuations or stock performance, based on a series of past events);

• Categorization analysis: a broad approach, bringing together parts of each method, which aims to explain the influence that many different factors have on a specific outcome.

(inspired from Info Strategy)

USA | Other Countries | NGO