Attitude trend analysis method and system applied to field of stock news

Abstract

The present invention discloses an attitude trend analysis method applied to the field of stock news. The method comprises: according to an attitude trend attribute of a stock news headline, establishing an initial training sample set composed of stock news headlines; preprocessing each stock news headline in the initial training sample set to obtain a processed training sample set; establishing a multi-layer LSTM neural network made of an input layer, an intermediate layer and an output layer, and training the neural network by using the processed training sample set, to obtain a trained neural network model; and determining the attitude trend attribute of a to-be-detected stock news headline by using the trained neural network model, to obtain the attitude trend attribute of the to-be-detected stock news headline. The method provided by the present invention realizes automatic analysis of the attitude trend of massive stock news, improves efficiency of labeling, increases accuracy of labeling, and has great significance to assisting investment decision-making. The present invention further provides a corresponding attitude trend analysis system applied to the field of stock news.

Description

technical field [0001] The invention belongs to the technical field of Internet finance, and more specifically relates to an attitude tendency analysis method and system applied in the field of stock news. Background technique [0002] With the rapid development of the domestic securities market, China's securities investors have exceeded 100 million, and 90% of investors are small and medium-sized retail investors. Market investment decisions dominated by retail investors are often significantly influenced by news public opinion and institutional research reports. Therefore, automatic analysis of the attitude tendencies of massive stock news is of great significance for assisting investment decisions. [0003] At present, the traditional labeling of news attitudes in the stock field is done manually, which is inefficient and insufficient to cover the increasing number of stock news in the era of information explosion. There have been some attempts in the industry to automa...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the above defects or improvement needs of the prior art, the present invention provides a method and system for analyzing attitude tendencies applied to the field of stock news, the purpose of which is to train a multi-layer LSTM (LongShort-Term Memory, long short-term memory network) neural network, obtain the trained neural network model, use the trained neural network model to discriminate the attitude tendency attribute of the stock news title to be detected, thereby solving the low efficiency of manual labeling and the high error rate of automatic labeling in the prior art technical issues

Images