Text normalization algorithm for facebook chats in hausa language

The 5th International Conference on Information and Communication Technology for The Muslim World (ICT4M)

The rapid increase in using non-standard words (NSWs) in communication through the social media is causing difficulties in understanding contents of the text messages. In addition, it affects the performance of several natural language processing (NLP) task such as machine translation, information retrievals, summarization and etc. In this study, we present an automatic text normalization system on Facebook chatting based on Hausa language. The proposed algorithm manually developed dictionary that employ normalization of each non-standard word with its equivalent standard word. This is accomplished through modification of the technique employed by [1] to fit Hausa NSWs' formation. It was found that our proposed algorithm was able to normalized Hausa NSWs with an accuracy of 100%The results of this research can facilitate comprehensive communication via Facebook using Hausa language.