nltk/nltk

How to stabilize NaiveBayesClassifier outputs?

Open

#2097 opened on Aug 28, 2018

View on GitHub
 (7 comments) (0 reactions) (0 assignees)Python (12,712 stars) (2,826 forks)batch import
good first issuetaggertests

Description

On sometimes we see that the continuous integration test fails because of some non-deterministic results from the NgramTagger i.e. are NaiveBayesClassifer, e.g. https://nltk.ci.cloudbees.com/job/pull_request_tests/822/PYV=3.6.4,jdk=jdk8latestOnlineInstall/testReport/junit/(root)/nltk/tag/

Is there someway to set a random seed or something such that classifier's output is consistent and make sure that the model is easily replicable?

Contributor guide

Tech stack
python
Domain
testingmachine learning
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-3 hours
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
Python basicsNLTK familiarity
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
70
Research direction
The issue requests adding a random seed parameter to NaiveBayesClassifier (in nltk/classify/naivebayes.py) or NgramTagger (in nltk/tag/ngram.py) to ensure reproducible outputs. A contributor should explore how other NLTK classifiers handle random seeds and implement a seed initialization using Python's random.seed() or numpy.random.seed() if applicable. The fix should be localized and include a test for reproducibility.