Text sentiment classification method based on deep learning combined model

Abstract

The invention discloses a text sentiment classification method based on a deep learning combined model. The method comprises the steps of (1) training a word vector and a character vector; (2) performing word segmentation for each sentence of an annotated text, filling to a fixed length to obtain a training dataset I, performing character segmentation on each sentence of the annotated text, filling to the fixed length to obtain a training dataset II; (3) endowing words and characters of the two training datasets with corresponding word vectors and character vectors; (4) training the two training datasets by use of a text CNN (Convolutional Neural Network) and an improved Dynamic CNN model to obtain four base classifiers, wherein a progressive learning method is adopted as a training methodand a focal loss function is adopted as a loss function and (5) combining the four base classifiers to obtain the text sentiment classification method of a combined model. The text sentiment classification method implemented by the invention does not depend on a specific sentiment dictionary, is not limited to a specific field and has high universality and expandability.

Description

technical field [0001] The present invention relates to the fields of deep learning and natural language processing, in particular to a text sentiment classification method based on a deep learning combination model. Background technique [0002] Text sentiment classification technology, its formal expression is: For a specific text x, the model predicts that the probability vector belonging to each sentiment category is P, and its category is: [0003] [0004] The main mainstream method of traditional text sentiment classification research is the classification method based on the sentiment lexicon. Usually, a lexicon containing various emotional colors is established first, and then the semantic information of the text is weighted based on the established sentiment lexicon, and then traditional machine learning methods for sentiment classification. The more commonly used emotional dictionaries include HowNet, NTUSD, and the Chinese emotional vocabulary ontology librar...

AI Technical Summary

Problems solved by technology

[0006] In the process of deep learning, model training is often easy to fall into local optimum or encounter saddle point problems, and the more layers, the more local optimum values ​​and saddle points exist in the neural network. At present, most of the schemes start from changing the weight initialization and propose Use Xavier, MSRA and other methods to initialize weights, but the effect is still very limited, and the training time is longer

Images