Overview
I built Tiny Language Classifier to test how well a simple Naive Bayes model could identify the language of a text.
It is a small experiment, but that is part of the point. A simple task does not always need a complex model.
Project details
I trained and tested the model on a language identification dataset from Hugging Face. The text is vectorized and passed to a Naive Bayes (MultinomialNB) classifier implemented with Scikit-Learn.
The model achieved over 90% accuracy on the test set.
Technologies used
- Python
- Scikit-Learn
- NumPy
- Pandas
- Hugging Face (datasets)