Improving Nyishi speech digit recognition with advanced cuckoo search optimization

Nyishi, a native tongue spoken by the Nyishi community in Northeast India, presents substantial hurdles in speech identification due to its distinct phonetic and syntactic features. Traditional speech recognition algorithms frequently fail to handle Nyishi correctly, resulting in misinterpretations and loss of linguistic complexity. This study provides an innovative method for boosting Nyishi speech digit recognition skills through Cuckoo Search Optimization (CSO). CSO, named after the brood parasitism behavior of cuckoo birds, is a powerful metaheuristic algorithm noted for its effectiveness in global optimization problems. By combining CSO with deep learning-based speech recognition models, our goal is to enhance the accuracy and resilience of Nyishi speech recognition technology. The suggested approach is tested on a newly compiled Nyishi speech dataset and shows considerable increases in recognition accuracy over conventional methods. The findings demonstrate the CSO’s ability to overcome the unique issues of Nyishi speech recognition, thereby contributing to the preservation and digital integration of this culturally important language. This study not only advances the field of speech recognition for low-resource languages, but also emphasizes the importance of optimizing algorithms in improving linguistic technologies.