Code-Mixed Sentiment Corpus of Customer Reviews in Malayalam and English for Enhanced Local Market Analysis in India

The popularity of online shopping has boosted the volume of customer reviews, which is crucial for business growth and decision-making as they influence consumers' purchasing decisions. Indians frequently engage in code-mixing, a social media practice where multilingual users publish comments that give information about consumer preferences, cultural relevance, and product or service improvement. Businesses with an Indian market emphasis should consider this factor to improve the consumer experience. Indian languages often face resource constraints, making it difficult to create precise language models and carry out operations like sentiment analysis, spam filtering, offensive text recognition, etc. Through this work, we draw the researcher's attention to the relevance of processing code-mixed customer reviews for e-commerce sites in linguistically diverse regions like India and how it helps to serve customers better, improve goods and services, and stay competitive in a globalised and multicultural market. This work aims to expand Malayalam-English (Manglish) code-mixed literature in the commercial domain by constructing a sentiment corpus featuring customer reviews in Manglish related to commercial products annotated by voluntary annotators. The performance of different machine learning (ML) and deep learning (DL) models for the sentiment identification task is then assessed using the new corpus, and a comprehensive analysis of the results is also presented.