News Text Classification Algorithm Based on Machine Learning Technology

Bo Zhang

DOI: https://doi.org/10.1109/icenit57306.2022.00047

2022-01-01

Abstract:With the rapid development of Internet technology, there are more and more text data on the Internet. Traditional manual text sorting methods can no longer cope with the current data volume. Automatic text sorting technology has become a research hotspot. As the main branch of text mining technology, text classification technology can effectively solve the needs of automatic text classification under the development of big data. The purpose of this article is to study news text classification algorithms based on machine learning technology. This article solves the problem that traditional machine learning methods require time and effort to manually construct mechanical features. This paper designs two common deep learning models, which can automatically extract feature convergence neural networks and long- and short-term memory neural networks. Experimental research shows that the classification effect of the topic model used in this article is better than that of the SVM model by 20%, which shows that it is advisable for us to select the topic model as a model for processing short text. Secondly, the result after adding the decision tree classification is also a certain improvement over the LDA model.