Financial news text emotional tendency analysis method based on graph convolutional network

Abstract

The invention discloses a financial news text emotional tendency analysis method based on a graph convolutional network. The method comprises the following steps: determining a data source to obtain financial text data; preprocessing the financial text data to obtain a clean text list; sampling the clean text list to obtain a sample list; carrying out manual labeling on the sample list; establishing a heterogeneous graph by using the clean text list; performing feature extraction on the heterogeneous image to obtain a feature matrix, a label matrix and an adjacent matrix; establishing a four-layer graph convolution network by taking the feature matrix as input, the label matrix as supervision information and the adjacent matrix as a support matrix of graph convolution operation; and obtaining the classification accuracy of the sample list and the classification result of the clean text list through iterative training. According to the method, the unlabeled data is introduced into the heterogeneous graph, learning can be carried out under the condition that no priori word embedding knowledge exists, and the dilemma that an emotion dictionary is difficult to construct and maintain in a web environment and the strong dependence on the proportion of labeled data and the word embedding effect are eliminated.

Description

technical field [0001] The invention relates to the technical field of natural language processing, in particular to a method for analyzing the emotional tendency of financial news texts based on a graph convolutional network. Background technique [0002] Text sentiment analysis is to give an evaluation of sentiment tendency to a given text through analysis. In the field of financial texts, sentiment analysis is used in financial crisis prediction research, financial news or comments can reflect the public's evaluation of listed companies; it is also used in investment analysis, financial news can better reflect market sentiment. At the same time, with the continuous development of information networks, the real-time, accurate and comprehensive coverage of web information also makes it possible to implement related tasks. [0003] Text sentiment analysis methods can be divided into two categories based on semantic analysis and machine learning methods. Among them, the met...

AI Technical Summary

Problems solved by technology

[0004] Among the above methods, the method based on semantic analysis needs to construct emotional dictionaries and semantic rules, which are very expensive to implement and maintain in the semantic environment of web information change; while many methods based on machine learning rely on word The effect of embedding, or requires a large amount of labeled data for training

Images