Profiling Engine SDK
Operator Reference

Operator Reference
Operator List
Topical List
   
 
Query Language
Introduction
Summary
Operators
Tips, Questions, and Answers
   
 
Main Index
Index
Tutorial
API Functions
Query Language
   
Technology Overview
   
Contact Us
   
 
Other Products
Onix Text Search and Retrieval Engine
Brevity Document Summarizer
Lextek Document Profiler & Categorizer
RouteX Document Routing Engine
Lextek Language Identifier
 

NORMALIZE

Name

NORMALIZE

Synopsis

Recomputes all hits for a term or subquery so that their weights fall between 0-1.

Arguments

Term or Subquery

Ranking Scheme

.

The weights are recalculated by dividing each hit's weight by the RMS (root mean square) average of all the weights for the term. This RMS method of normalizing values is often found in physics or engineering. You take the average of the square of each weight. You then take the square root of that average.

Any resultant weights which would be larger than 1 are truncated to 1.

 

Comments

NORMALIZE is designed to help you put weights between 0 and 1. Typically this is done as the last step of your query when you have used weights that were larger than 1. This can occur when biasing terms to make them more relevant than terms which already had a weight of 1.

 

Example

NORMALIZE( 'blue' );

See Also

maxnormalize, mynormalize, complement