A Classification Method for Identification of Public Opinion Tendency Aiming at Unbalanced Category Distribution

Abstract

The invention discloses a public opinion tendency recognition method aiming at the unbalanced distribution of training sample categories. First, collect words related to the field of public opinion concerned as hot words of public opinion to create a thesaurus; crawl the comment data set from the source of public opinion, and divide it into a training set and a test set. Then, the public opinion tendency of the training set is manually classified, and the bootstrap learning method is used to supplement the problem of category imbalance. Extract the characteristics of each type of training sample, use Naive Bayesian, support vector machine, decision tree and other algorithms to train the algorithm model, use the trained model to classify the test set data, and identify the tendency of public opinion according to the classification results. The methods of bootstrap learning, eigenvector construction and classification model training are all weighted by time-sensitive weighting methods, so that the public opinion tendency reflected by them is more time-sensitive. The invention solves the problem of inaccurate classification caused by unbalanced training data, and improves the accuracy of public opinion tendency recognition and the timeliness of public opinion analysis.

Description

technical field [0001] The invention belongs to the technical field of natural language processing and machine learning, and relates to a method for analyzing the tendency of public opinion by using a machine learning algorithm, in particular to a method for identifying the tendency of public opinion aimed at the unbalanced distribution of training sample categories. Background technique [0002] The current Internet penetration rate is growing rapidly, the number of updated news on the Internet is very large, and the resulting impact on public opinion is also very large. It is under this situation that the analysis technology of public opinion tendency was born, aiming at analyzing the public opinion generated on the Internet. Reviewers' tendentious attitudes and attitude changes are screened in a timely manner, thereby helping regulatory authorities to detect changes in public opinion in a timely manner and build a civilized and harmonious public opinion environment. [00...

AI Technical Summary

Problems solved by technology

[0003] When using general-purpose machine learning algorithms to analyze public opinion tendencies, problems such as class imbalance of training data, timeliness of text release, and timeliness of public opinion often lead to a large deviation between the effect of tendency recognition and the actual tendency

Currently, no effective solution has been proposed

Images