CachingNaiveBayesClassifier doesn't compute log prior. [LUCENE-7672]
- Dominant language
- Java
- Stars
- 3.6k
- Forks
- 1.4k
- Avg merge
- 2d 11h
- Merged PRs (30d)
- 88
Description
The CachingNaiveBayesClassifier gives different results than the SimpleNaiveBayesClassifier. This is due to the fact that the CachineNaiveBayesClassifier does not calculate the log prior on the class (category) names, as it should.
This bug has been fixed, and a patch has been attached to this issue.
This affects master, as well as previous releases up to version 6.4.
---
Migrated from [LUCENE-7672](https://issues.apache.org/jira/browse/LUCENE-7672) by Kevin Crosby, updated Jul 19 2020
Attachments: [LUCENE-7672.patch](https://apache.github.io/lucene-jira-archive/attachments/LUCENE-7672/LUCENE-7672.patch)
Contributor guide
Research direction
Start by locating CachingNaiveBayesClassifier and SimpleNaiveBayesClassifier, then compare how they calculate class or category priors. Review the attached LUCENE-7672.patch for the intended fix and verify that both classifiers produce matching results for the affected cases.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100