AI::Ca...er::KNN(3pm)

💖 AI::Ca...er::KNN — wait, it's AI::Categorizer::Learner::KNN 💖

okay bestie so this is the User Contributed Perl Documentation page for AI::Categorizer::Learner::KNN, aka the K Nearest Neighbour algorithm doing its thing for AI::Categorizer. yes the title is truncated, no we're not gonna let that stop us. sit down, hydrate, let's go 🫶

---

✨ NAME ✨

AI::Categorizer::Learner::KNN — the K Nearest Neighbour Algorithm for AI::Categorizer. that's it. that's the name. iconic.

---

🌸 SYNOPSIS 🌸

here's the whole vibe, verbatim, do not touch the code bestie:

  use AI::Categorizer::Learner::KNN;

  # Here $k is an AI::Categorizer::KnowledgeSet object

  my $nb = new AI::Categorizer::Learner::KNN(...parameters...);
  $nb->train(knowledge_set => $k);
  $nb->save_state('filename');

  ... time passes ...

  $l = AI::Categorizer::Learner->restore_state('filename');
  my $c = new AI::Categorizer::Collection::Files( path => ... );
  while (my $document = $c->next) {
    my $hypothesis = $l->categorize($document);
    print "Best assigned category: ", $hypothesis->best_category, "\n";
    print "All assigned categories: ", join(', ', $hypothesis->categories), "\n";
  }

and yes $nb is the variable name even though this is KNN and not Naive Bayes. we don't question the ancestors. we respect the ancestors. 🙈

---

🦋 DESCRIPTION 🦋

so this module is a whole implementation of the k-Nearest-Neighbor decision-making algorithm, aimed at the task of document categorization (as defined by the AI::Categorizer module). the entire interface is spelled out in AI::Categorizer, so go read that for the full picture. this page is the KNN-specific tea. 💅

---

💅 METHODS 💅

this class inherits from the AI::Categorizer::Learner class, so all of its methods are available to you unless it's explicitly mentioned here. inherited methods, we love to see it. 🎀

✨ new() ✨

creates a fresh new KNN Learner and hands it back to you. on top of the parameters that the AI::Categorizer::Learner class already accepts, the KNN subclass accepts these extra ones:

• threshold — sets the score threshold for category membership. the default is currently 0.1. want more categories per document? set the threshold lower. want fewer? set it higher. it's a lowkey effective way to trade off between precision and recall. balance queen. ⚖️
• k_value — sets the "k" value (as in k-Nearest-Neighbor, yes the name is right there) to the given integer. this tells it how many of each document's nearest neighbors get considered when assigning categories. default is 5. 👑

✨ threshold() ✨

returns the current threshold value. and if you pass it an optional numeric argument, you can set the threshold too. getter and setter, main character energy fr fr. 🌟

✨ train(knowledge_set => $k) ✨

trains the categorizer. this is the prep step so it can actually categorize documents later. the knowledge_set parameter has to provide an object of the class AI::Categorizer::KnowledgeSet (or a subclass of it), populated with lots of documents and categories. for the whole how-do-i-make-one-of-those situation, check AI::Categorizer::KnowledgeSet. 📚

✨ categorize($document) ✨

returns an AI::Categorizer::Hypothesis object representing the categorizer's "best guess" about which categories the given document should be assigned to. want the details on how to actually use that object? AI::Categorizer::Hypothesis has you covered. 🔮

✨ save_state($path) ✨

saves the categorizer so you can use it later. heads up bestie, this method is inherited from AI::Categorizer::Storable. it's not even KNN's own method, it's just vibing here. 💾

---

🫶 AUTHOR 🫶

originally written by David Bell (<dave@student.usyd.edu.au>), October 2002. 🎂

then it was added to AI::Categorizer in November 2002, modified, and maintained by Ken Williams (<ken@mathforum.org>). legends only. 👑

---

💜 COPYRIGHT 💜

Copyright 2000-2003 Ken Williams. All rights reserved.

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself. slay, open source. 🌈

---

🌟 SEE ALSO 🌟

• AI::Categorizer(3)
• "A re-examination of text categorization methods" by Yiming Yang — <http://www.cs.cmu.edu/~yiming/publications.html> — yes it's a link, yes it's in the original, yes you should read it if you're into the theory. academic baddie behaviour. 📖

---

ok that's the whole page bestie, every flag, every default, every method, nothing skipped 💖

and that's a wrap! this was perl v5.42.0, dated 2025-07-25, straight out of the User Contributed Perl Documentation collection. go categorize something, you've got this. periodt. 💅✨