Christian Dave O. Empinado, Robine Cole A. Jumalon, Christine F. Peña
Code-Switching (CS) is a phenomenon where a person switches from one language to another in a single conversation. In the Philippines, CS is used often due to multiple languages that coexist together in a society as it is a simpler way of conveying messages between its members. This study will introduce a method that detects code-switching, specifically in the Cebuano-English language. A support vector machine (SVM) with language parameters from a Gaussian Mixture Model (GMM) was used for language classification. An Automatic Speech Recognition (ASR) system, consisting of Cebuano acoustic and language models, was developed to aid in the detection process. By combining language classification information and acoustic score, the method showed an increased accuracy of code-switching speech recognition compared to previous methods. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
University of San Carlos, P. del Rosario St, Cebu City, 6000, Philippines
