Add MultinomialNaiveBayesClassifier Implementation

[Rosander0]

Add Multinomial Naive Bayes classifier

Adds a MultinomialNaiveBayesClassifier implementation to com.thealgorithms.machinelearning, following the same package structure as the existing Linear Regression implementation.

What's included:

Multinomial Naive Bayes with Laplace (add-alpha) smoothing, configurable via constructor (defaults to alpha = 1.0)
Log-probability arithmetic throughout predict() to avoid numerical underflow from multiplying small probabilities directly
Single-sample and batch predict() overloads
Guard clauses for unfitted-classifier use, invalid alpha, and feature/sample dimension mismatches
JUnit 5 test suite covering prediction correctness on a separable toy dataset, batch/single-sample consistency, smoothing behavior on zero-count features, and all exception paths

Notes for reviewers:

This is the Multinomial variant specifically (suited to count-based features like word frequencies), not Gaussian or Bernoulli — happy to add those as follow-ups if there's interest
One test (laplaceSmoothingKeepsZeroCountFeatureLogProbabilityFinite) only asserts the prediction doesn't collapse to an invalid state, not the exact expected class — flagging in case a maintainer wants that tightened

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 80.40%. Comparing base (38ff681) to head (c1699f0).

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #7532      +/-   ##
============================================
+ Coverage     80.35%   80.40%   +0.04%     
- Complexity     7425     7444      +19     
============================================
  Files           813      814       +1     
  Lines         23955    24016      +61     
  Branches       4717     4727      +10     
============================================
+ Hits          19250    19309      +59     
- Misses         3944     3945       +1     
- Partials        761      762       +1     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Rosander0]

Hey @DenizAltunkapan , just checking in on this PR — it's been a few days. Happy to make any changes if something needs to be adjusted. Let me know!

[DenizAltunkapan]

@Rosander0 Thanks for the contribution. The implementation looks solid overall, but fit() should clear logPriors and logLikelihoods before training. Otherwise, calling fit() a second time can leave classes from the previous training run in the model and affect later predictions.

Please also add a regression test for refitting the same classifier instance. After that, this should be ready to approve.

[DenizAltunkapan]

@Rosander0 thanks!