Research On Trends Of Steel Price Index Based On Sentiment Analysis

China is the world's largest iron and steel producer.It is significant to predict the steel price indexes in practical application.The traditional prediction methods mainly use basic data,combined with mathematical statistics and machine learning methods to model,analyze and predict,with the need to be expanded in data sources.Based on the systemic study of sentiment analysis technology,upstream and downstream chains of steel industry and steel price index,this paper establishes a trend prediction framework of steel price index based on sentiment analysis,obtains unstructured texts such as steel industry news and research reports,and forms a corpus of sentiment analysis,and constructs a dictionary of inversion words,emotional dictionary,negative dictionary,special word dictionary,domain emotional object dictionary,domain emotional attribute dictionary and other iron and steel field emotional dictionary set;designed the upstream and downstream chain and research nodes of iron and steel industry;classified,selected,and focused on a variety of iron and steel price indexes;determined the prediction model and the corresponding data points.Finally,the trend change of steel industry index is predicted,and the effective and practical experimental results are obtained.