To extract effective information in massive financial news,this paper proposes amethod to calculate the correlation between text and text set by extracting structured events in the stockmarket news text,and provides m...To extract effective information in massive financial news,this paper proposes amethod to calculate the correlation between text and text set by extracting structured events in the stockmarket news text,and provides more detailed and interpretable information.First,the structured event triplet was extracted from the text set,and the trained word vector was used to represent the event triplet as an event vector.Event vectors were clustered,the cosine distances were calculated for the cluster centers,and the correlation between the text sets was determined by matching.Finally,the event triplets with the highest correlation between the text sets were selected to provide explanation information for the calculation results.Experimental results show that this method effectively measures the correlation between text and text set.展开更多
文摘To extract effective information in massive financial news,this paper proposes amethod to calculate the correlation between text and text set by extracting structured events in the stockmarket news text,and provides more detailed and interpretable information.First,the structured event triplet was extracted from the text set,and the trained word vector was used to represent the event triplet as an event vector.Event vectors were clustered,the cosine distances were calculated for the cluster centers,and the correlation between the text sets was determined by matching.Finally,the event triplets with the highest correlation between the text sets were selected to provide explanation information for the calculation results.Experimental results show that this method effectively measures the correlation between text and text set.
