Customers need to be able to talk to businesses in more than one language in the very competitive world of e-commerce. Sentiment analysis is now a vital tool for improving business efficiency and making smart choices. Previous studies on sentiment analysis focused on English, which made it harder to get accurate results from reviews written in English. To make the model more accurate, English reviews were added to a machinelearning model. In a thorough study, accuracy, precision, recall, and F1 scores were used to compare three algorithms. The dataset, which was discovered on Kaggle, was meticulously labeled with positive, negative, and neutral sentiments. After preprocessing the data, machine learning approaches were employed to train the model and evaluate its performance. The accuracy of Multinomial Naive Bayes (MNB) and Random Forest (RF) was 93%, while Decision Tree (DT) was 91%. It was critical to collect and annotate the dataset to ensure its quality and applicability for sentiment analysis algorithms. Experimenting with different algorithms and finetuning hyper parameters may show potential for improvement. For accurate and meaningful results, a broad and representative dataset is required. A greater range of products, sectors, and consumer demographics would boost generalizability. The study on sentiment analysis for English adds greatly to the discipline this research could help English-speaking companies comprehend consumer mood. E-commerce enterprises can better serve customers by improving sentiment analysis. This may provide these companies with a market advantage. The researchers expect the study will impact the e-commerce industry. The objective of this research is to improve sentiment analysis in the e-commerce industry, particularly for customer evaluations written in English. The researchers aim to improve the accuracy of sentiment analysis models by including English reviews in a machine-learning model.
Share this article