Deep Learning Based Model for a Spell Checker of Wolaita Language

Natural Language Processing (NLP), a subfield of Artificial Intelligence (AI), facilitates interactions between humans and computers to enable language understanding, with a crucial application like spell checking, which ensures word accuracy in large document collections. The Wolaita language widely used by government and non-government organizations for educational purposes, news, and university studies. Due to lack of an effective spell-checker tool for Wolaita language, resulting in difficulties during document preparation, reduced quality, and inefficiencies. To address this, the study aimed to design and develop a spell checker model for Wolaita, utilizing deep learning algorithms such as BiLSTM, LSTM, GRU, and RNN, combined with suitable feature extraction techniques. Experimental results revealed that the BiLSTM model achieved the highest accuracy at 98.12%, followed by LSTM at 95.97%, GRU at 91.35%, and RNN at 86.37%,demonstrating that BiLSTM is the most effective for Wolaita spell checking. The study concludes that BiLSTM is the most reliable model for developing a Wolaita spell checker, and future research should explore integrating additional language resources and larger datasets to improve accuracy and adaptability.