A Recurrent Neural Network Text Sentiment Analysis Method Embedded with Logical Rules

Abstract

The present invention provides a cyclic neural network text sentiment analysis method embedded with logic rules, by grabbing the text corpus used for training, carrying out emotion category marking, and then dividing the text corpus of emotion marking into training set corpus and test set corpus, and Perform word segmentation processing and stop word processing on it, and then use the word2vec algorithm to train the training set corpus and test set corpus after word segmentation processing and remove stop words, obtain corresponding word vectors, and combine the training set corpus and test set corpus Collecting corpus and inputting the existing knowledge base in combination with the probability graph model for analysis and processing, and embedding the first-order logic rules into the cyclic neural network through the logic cyclic neural network structure (Logic-RNN and Logic-LSTM), on the one hand, the present invention can achieve Controlling the training direction of the cyclic neural network is more inclined to human intuition. On the other hand, it improves the accuracy of text sentiment analysis. This method can also be used in other fields of natural language processing and machine learning.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a text sentiment analysis method for embedding logic rules in Recurrent Neural Networks (RNNs). Background technique [0002] With the development of Internet technology and the rise of web2.0, the Internet has gradually changed from a static information carrier to a platform for people to obtain information, express opinions, and exchange emotions. People share, comment, and express their opinions on various things online. , Opinions, such as comments on movies, news, stocks, etc. These comments are self-evident for the importance of the government, enterprises, consumers, etc. However, with the explosive growth of online comment data, relying on manual processing of massive text data Acquisition, processing, analysis, and prediction are impractical, so using automated tools to quickly obtain valuable information from a large amount of text has become an urgent need for ...

AI Technical Summary

Problems solved by technology

[0007] As a sequence model, the cyclic neural network has achieved great success and is widely used in many natural language processing tasks, such as language recognition, machine translation, sentiment analysis, entity recognition, etc., which makes more and more people believe in the cyclic neural network. It can be a good language model, but there are still many shortcomings in the cyclic neural network. For example, the training of the cyclic neural network takes a lot of time, and the high-precision model depends on a large amount of data. Simple data learning often leads to inexplicable sex and counterintuitive

Images