Speech and NLP Models for Business Communication Automation
Abstract
Natural Language Processing (NLP) is increasingly used in business communication to enhance customer experience, employee engagement, and brand interaction. This paper evaluates the ability of two modelsLogistic Regression and Long Short-Term Memory (LSTM)to classify emotions in text-based communication. Using a labelled dataset of six emotion categories (joy, sadness, anger, fear, surprise, and love), the models were trained, validated, and tested. Quantitative results show that while Logistic Regression achieves competitive baseline performance (accuracy 87.7%), LSTM outperforms it $(92.4 \%)$ due to its ability to capture sequential dependencies and contextual meaning. The findings highlight the practical relevance of emotion-aware NLP systems in real-time business communication automation.