BemMorph

June 1, 2024

Building a Morphological Analyzer for Bemba

As part of my final year dissertation, I developed a morphological analyzer for Bemba, a Bantu language spoken in Zambia. It is build using a Finite State Transducer and my implementation heavily relies on the [Foma FST library](https://fomafst.github.io/. You can explore my project a:

• bemmorph.martinnn.com.

My Implementation

I set up the morphological analyzer using a Flask server. The routing is managed with Nginx and secured with Let’s Encrypt for SSL.

Features

• Partial Morphological Analysis: Breaks down Bemba words into their constituent morphemes.
• Edit Distance-Based Word Search: Helps users find words similar to a given input using edit distance.

Current Status

The project is currently in an incomplete state, but I plan to continue updating and improving it. The main issue at the moment has to do with overgeneration and minimal word coverage(current coverage mostly limited to verbs).

Why It’s Useful

Morphological analyzers are valuable tools for linguists, language learners, and researchers. They help in understanding the structure of words and exploring word variations across different languages. For more information on morphological analysis, you can refer to this Wikipedia article.

For more details, visit my morphological analyzer.