Unveiling the Emotions: A Sentiment Analysis of Amazon Customer Feedback

This study explores sentiment analysis in the context of diverse regions and contemporary customer feedback, aiming to address research questions related to consolidation based on polarity scores and sentiments. The research utilizes multinomial regression for a comprehensive analysis of customer feedback worldwide. The investigation incorporates confusion matrices, statistics, and class-specific metrics to evaluate the model’s performance. Results indicate a highly accurate model with perfect sensitivity, specificity, and overall accuracy. The analysis further includes a breakdown of key metrics such as accuracy, confidence intervals, no information rate, p-value, kappa, and prevalence, emphasizing the model’s robustness. In conclusion, the multinomial logistic regression model demonstrates exceptional performance in predicting sentiment across diverse classes, highlighting its effectiveness in sentiment analysis on a global scale.