MaCmS: Magahi Code-mixed Dataset for Sentiment Analysis
This addresses the lack of resources for sentiment analysis in Magahi, a less-resourced minority language, but is incremental as it primarily creates a new dataset.
The paper introduces MaCMS, the first Magahi-Hindi-English code-mixed dataset for sentiment analysis, and provides linguistic and statistical analyses along with baseline model evaluations.
The present paper introduces new sentiment data, MaCMS, for Magahi-Hindi-English (MHE) code-mixed language, where Magahi is a less-resourced minority language. This dataset is the first Magahi-Hindi-English code-mixed dataset for sentiment analysis tasks. Further, we also provide a linguistics analysis of the dataset to understand the structure of code-mixing and a statistical study to understand the language preferences of speakers with different polarities. With these analyses, we also train baseline models to evaluate the dataset's quality.