Research on encrypted text classification based on natural language processing

Abstract

In reality, data encryption technology is mostly used to protect the security of text data in the network, but when we need to obtain these data, this layer of encryption becomes an obstruction to obtaining data. The general method uses data mining and data decryption to extract effective information. The experimental data in this article selected 20 categories of text information, and obtained a data set with a difficulty of 1 to classify the encrypted text information. In order to classify encrypted text more effectively, this paper studies the method of using the logistic regression model and the LightGBM model algorithm to directly process encrypted text, which can directly extract and classify the text in the encrypted state. Model evaluation results show that LightGBM is more effective. In addition, this article provides a basic framework for the classification of encrypted text based on natural language processing.