A Stochastic Collocation Algorithm Method for Processing the Yoruba Language Using the Data Context Approach Based on Text, Lexicon, and Grammar

摘要

In this paper, we show an initial attempt to generate a self-extractive text processor for the Yoruba language. The Yoruba language is a language spoken by about 60 million persons across America, Europe, and majorly West Africa. This is implemented with the use of a holder codenamed "YOTEX". YoTEx is a Yoruba language text repository which simply learns from the English Language corpus with much emphasis on the agglutinative tendencies of the Yoruba language.In the building of the data repository, the development of the system considered parameters for existing relations as available in other textual corpora like the WordNet English corpus, which is used in this work as a case study. We used stochastic collocation algorithm to show relationship within entities. The choice of the algorithm is based on the tonal orientation of the language. Hidden Markov model was extended in line with the aim of carrying out deep text analysis. The developed system performs well against known benchmarks in the formulation of an appropriate tagging, part of speech, stemming, chunking etc. system for the Yoruba textual terms. The resulting YoTex will improve the "codinazation" of the Yoruba Language in particular and the other agglutinative language in general. Such will enhance the computer processing efficacies of the Yoruba language. This work presents a novel approach of testing some known language models on a Yoruba lexical corpus.