Using a Small Lexicon with CRFs Confidence Measure to Improve POS Tagging Accuracy

Using a Small Lexicon with CRFs Confidence Measure to Improve POS Tagging Accuracy Mohamed Outahajala1, Paolo Rosso2 1Institut Royal de la Culture Amazighe, Avenue Allal El Fassi Madinat Al Irfane - Rabat - Instituts Adresse postale : BP 2055 Hay Riad Rabat, Morocco 2 NLE Lab, PRHLT Research Center, Universitat Politècnica de València, Spain E-mail: outahajala1@yahoo.fr , prosso@dsic.upv.es Abstract Like most of the languages which have only recently started being investigated for the Natural Language Processing (NLP) tasks, Amazigh lacks annotated corpora and tools and still suffers from the scarcity of linguistic tools and resources. The main aim of this paper is to present a new part-of-speech (POS) tagger based on a new Amazigh tag set (AMTS) composed of 28 tags. In line with our goal we have trained Conditional Random Fields (CRFs) to build a POS tagger for the Amazigh language. We have used the 10-fold technique to evaluate and validate our approach. The CRFs 10 folds average level is 87.95% and the best fold level result is 91.18%. In order to improve this result, we have gathered a set of about 8k words with their POS tags. The collected lexicon was used with CRFs confidence measure in order to have a more accurate POS-tagger. Hence, we have obtained a better performance of 93.82%. Keywords: POS tagging, CRFs, lexicons. 1. Introduction Amazigh is spoken in Morocco, Algeria, Tunisia, Libya, and the Egyptian Oasis Siwa; it is also spoken by many other communities in parts of Niger and Mali and by immigrant Amazigh communities in Europe and over the world. Amazigh language belongs to the Hamito-Semitic languages (Cohen 2007, Chaker 1989, Chafiq 1991) with rich templatic morphology. With the emergence of an increasing sense of identity, Amazigh speakers would very much like to see their language and culture rich and developed. In Morocco, Amazigh has been introduced in mass media and in the educational system in collaboration with relevant ministries. Accordingly, the first and the second Moroccan channels begun to broadcast some programs in the Amazigh language in 2007. A new Amazigh television channel was launched in the first of March, 2010. It has also become common practice to find Amazigh taught in various Moroccan schools as a subject. On July 2011, Moroccans voted favourably for the new constitution; therefore, the Amazigh language became an official language along with Arabic. In February 2016, Amazigh language was made an official language by the Algerian government. However, this language, like most of the languages which have only recently started being investigated for NLP, still suffers from the scarcity of language processing tools and resources. In this sense, since POS tagging is an important and basic step in the processing of any given language, the main objective of this paper is to explain how we improved an Amazigh POS tagger accuracy. The rest of this paper is organized as follows: in Section 2 we present related works about Amazigh POS tagging. Then we describe the more fine-grained tag set for Amazigh. The fourth …

Content truncated to 500 words