This site has been initiated to serve the Amharic NLP and CL community of interest.

This site is also intended serve as a starting point for anyone working on natural language processing and computational linguistics for the Amharic language.

New Resources

AMMORPHO is a Python program that analyzes Amharic words into their constituent morphemes (meaningful parts), returning, whenever possible, the stem or root of the word, along with a representation of its grammatical structure. It is the first in a planned series of programs to handle the morphology of Ethiopian and Eritrean languages, and some of the decisions have been made in order to maintain consistency across languages, especially those within the Semitic family.

New Papers

Semitic Morphological Analysis and Generation Using Finite State Transducers with Feature Structures, Michael Gasser, Indiana University, 2008

Amharic Speech Recognition: Past, Present and Future, Solomon Teferra Abate, Martha Yifiru Tachbelie, Wolfgang Menzel, 2007

Syllable Based Speech Recognition for Amharic, Solomon Teferra Abate, Wolfgang Menzel, 2007

Automatic Speech Recognition for an Under-Resourced Language - Amharic, Solomon Teferra Abate, Wolfgang Menzel, 2007

Automatic Classification of Amharic News Items: The Case of the Ethiopian News Agency, Zelalem Sintayehu, MS Thesis, 2001

Manual Annotation of Amharic News Items with Part-of-Speech Tags and its Challenges, Girma A. Demeke & Mesfin Getachew, ELRC Working Papers Vol. 2 No 1, March, 2006

Non-concatinative Finite-State Morphotactics of Amharic Simple Verbs, Saba Amsalu & Girma A. Demeke, ELRC Working Papers Vol. 2 No 3, September, 2006