Classification of Sentiments of the Roman Urdu Reviews of Daraz Products using Natural Language Processing Approach

Sentiment Analysis plays a significant part in determining how the audience feels about a specific product or topic. It is an essential fragment in the machine learning and natural language processing domain. In this paper, sentiment analysis of the reviews was performed. The dataset was obtained from Kaggle containing the reviews of different products in Roman Urdu text posted on Daraz E-Commerce site. The data is collected and then processed. The finalized dataset is then analyzed for the classification of the sentiments using different machine learning classifiers in Python. The results and accuracies reported by the algorithms are also presented and discussed in detail in the paper. Logistic Regression outperforms with an accuracy of 75%