BERT-Enhanced Deep Gaussian Regression: A Unified Approach for Sentiment Analysis and Price Prediction

Abstract

This study presents a novel approach that fuses the BERT (Bidirectional Encoder Representations from Transformers) model for sentiment analysis with Deep Gaussian Processes (DGP) for automotive spare parts price prediction, targeting applications in supply chain management. Leveraging the strengths of both models, our method enhances NLP (Natural Language Processing) and time series forecasting, promoting informed decision-making and improved business intelligence strategies in a data-driven world.