Improving Alexa’s response with sentiment analysis and comparing various machine learning models

This research paper presents a comparative analysis of four machine learning models: Random Forest, Decision Tree Classifier, XG Boost, and Artificial Neural Networks (ANNs), evaluating their performance using various metrics such as accuracy, precision, recall, F1 scores, and log loss. Through rigorous testing, ANNs demonstrated superior generalization capabilities, exhibiting the highest testing accuracy and the lowest log loss, indicating a solid ability to handle unseen data with high predictive confidence. Conversely, while Random Forest and Decision Tree classifiers showed excellent training accuracy, they were less effective in testing scenarios, suggesting limitations in their generalization potential. XG Boost performed well across multiple metrics, closely competing with ANNs. The study employs confusion matrices, ROC curves, and other visual tools to provide a comprehensive overview of each model’s strengths and weaknesses, guiding the selection of the most appropriate model based on specific requirements and the nature of the predictive task. This analysis is crucial for applications that demand high reliability and accuracy in predictions on new datasets, highlighting ANNs as particularly suitable for such tasks.