Computing load issues

From RapidWiki

Jump to: navigation, search

This article is a sketch. Feel free to modify it

Contents

Memory settings

Java virtual memory
Amount of memory
Parameters to be changed in two files
Creating new RapidMiner environment variables

Software inner memory management

Sparse data formats
"Database one-line" mode.

Choice of the right operator

Example : FPGrowth vs A PRIORI
Example : Feature Selection before reduction, and reduction before classification.

Testbench and examples

Table :
Type of algorithm
Type of machine
Time measured

Projects

Development makes RapidMiner faster and less greedy.
"Cluster machine" mode ?