A Comparative Study for Language Recognition using Learning-based Approaches

Language recognition is helpful for determining the natural language in a given document or part of text. Language recognition has attracted more attention in recent times due to its wide-ranging applications, including speech translation, multilingual speech recognition and more. Indeed, language recognition should be effective to ensure practical implementation. Therefore, learning-based approaches are introduced to enhance the effectiveness of language recognition. In this paper, a total of six learning-based approaches have been implemented for solving the language recognition problem. Experiments and evaluations are conducted to study the effectiveness of these learning-based approaches on identifying 5 different languages which are English, German, Czech, French, and Swedish. The experimental results show that the 1D-CNN model achieves the highest accuracy score of 65.99%.