---
name: "AI::DecisionTree::Instance"
section: "3pm"
platform: "linux"
package: "libai-decisiontree-perl"
version: "0.11-2+b4"
release: "Debian 13 (Trixie)"
platforms: ["linux"]
url: "https://new.carta.tech/AI::DecisionTree::Instance"
---

# AI::DecisionTree::Instance(3pm)

C-struct wrapper for training instances

## NAME

AI::DecisionTree::Instance - C-struct wrapper for training instances

## SYNOPSIS

```
  use AI::DecisionTree::Instance;

  my $i = new AI::DecisionTree::Instance([3,5], 7, 'this_instance');
  $i->value_int(0) == 3;
  $i->value_int(1) == 5;
  $i->result_int == 7;
```

## DESCRIPTION

This class is just a simple Perl wrapper around a C struct embodying a single training instance.  Its purpose is to reduce memory usage.  In a "typical" training set with about 1000 instances, memory usage can be reduced by about a factor of 5 (from 43.7M to 8.2M in my test program).

A fairly tight loop is also implemented that helps speed up the `train()` AI::DecisionTree method by about a constant factor of 4.

Please do not consider this interface stable - I change it whenever I have a new need in AI::DecisionTree.

## AUTHOR

Ken Williams, ken@mathforum.org

## SEE ALSO

AI::DecisionTree