Replication Data for NSEformer: News-Signal Extractor Transformer for Carbon Price Prediction

Due to fierce problems caused by global warming, more and more countries have implemented various carbon emission restrictions and emissions trading schemes(ETSs) around the world, leading to emergence of global carbon trading markets and a need for accurate carbon price predictions. Our approach enhances Transformer-based time-series models by incorporating news data as strength indicators for price increases. We propose NSEformer, a novel way to combine text data into time-series models using News-Signal Extractor. Through experiments, NSEformer outperforms other methods, offering superior performance in predicting carbon trading prices. Our approach effectively integrates news information, leveraging the expertise of the News-Singal Extractor, and outperforms other fusion strategies.