A Machine Learning Model for the Dating of Ancient Chinese Texts

Abstract

This paper, with the intent of solving the issues on the dating of ancient Chinese texts, takes advantage of the Long-Short Term Memory Network (LSTM) to analyze and process the character sequence in ancient Chinese. In this model, each character is transformed into a high-dimensional vector, and then vectors and the non-linear relationships among them are read and analyzed by LSTM, which finally achieve the dating tags. Experimental results show that the LSTM has a strong ability to date the ancient texts, and the precision reaches about 95% in our experiments. Thus, the proposed model offers an effective method on how to date the ancient Chinese texts. It also inspires us to actively improve the time-consuming analysis tasks in the Chinese NLP field.