Background of the Study
Language modeling is a fundamental aspect of natural language processing, designed to predict word sequences and capture linguistic patterns. In Nigeria, language use is often characterized by code-switching between Nigerian English and Pidgin. This code-switched speech is prevalent in social media, informal communication, and even formal contexts. Recent advancements in deep learning have enabled the development of sophisticated language models that can handle multilingual and code-switched data (Adejumo, 2023). However, the dynamic nature of code-switching, with its frequent alternation between languages and the blending of grammatical structures, poses unique challenges for conventional language models. Studies (Okeke, 2024) suggest that models trained on monolingual data often fail to capture the nuances of code-switched communication. Recent research (Umar, 2025) emphasizes the need for specialized training on mixed-language corpora to improve model performance. This study examines language modeling approaches for code-switched Nigerian English and Pidgin corpora, assessing model accuracy and proposing adaptations that account for the complexities of bilingual language use in Nigeria.
Statement of the Problem
Existing language models frequently struggle with code-switched text, leading to poor prediction accuracy and limited applicability in real-world Nigerian contexts (Adejumo, 2023). The alternating structures and blended vocabularies in Nigerian English and Pidgin create challenges that monolingual models are not equipped to handle (Okeke, 2024). This deficiency impacts applications such as speech recognition, translation, and sentiment analysis. Furthermore, the scarcity of annotated code-switched corpora exacerbates the problem. Addressing these issues is essential to develop language models that accurately reflect the bilingual dynamics of Nigerian communication.
Objectives of the Study
Research Questions
Significance of the Study
This study is significant because it addresses the complex issue of code-switching in Nigerian language use by evaluating and enhancing language models. Improved models will benefit applications such as speech recognition, machine translation, and sentiment analysis, leading to more accurate digital tools for multilingual communities. The findings will provide insights for researchers and developers working on bilingual NLP systems and contribute to the preservation and effective processing of Nigeria’s linguistic diversity.
Scope and Limitations of the Study
This study focuses on language modeling for code-switched Nigerian English and Pidgin corpora and does not cover other language pairs or broader sociolinguistic factors.
Definitions of Terms
ABSTRACT
The main purpose of this research work was to assess the effects of stress on the academic performance of stude...
Background of the Study
In the competitive banking sector, effective cost control is essential to ensuring branch profitabi...
Background to the Study
The importance of marital adjustment cannot be over-emphasized; the adjustment of each marriage...
Background of the Study
Physical fitness is a fundamental aspect of overall well-being, contributing not only to physical...
Background of the Study
Rural banking reforms are critical for improving the accessibility of agricultural...
Background of the Study
Nursing workload is a critical factor influencing the quality of patient care, particularly in p...
Background of the Study
Rural electrification has become a key driver of socioeconomic development in remote areas, significantly impacti...
Background of the Study
Since becoming the Minister of the Federal Capital Territory (FCT) in 2023, Nyesom Wike has been ta...
BACKGROUND TO THE STUDY
“The more outré and grotesque an incident is the more carefully it...
Background of the Study
Nutrition plays a crucial role in ensuring the health and well-being of pregnant women and their unborn children....