Artistic Conception Analysis Of Classical Poetry Based On Deep Learning

Chinese classical poetry is a treasure of Chinese culture,besides,it also contains the spirit of the Chinese nation.It is the crystallization of the wisdom of the ancient Chinese people.Learning Chinese ancient poetry can not only understand history and culture,enhance national pride,but also enrich people’s spiritual world.At present,most of the research work on ancient poetry relies on the human research of literature-related workers.Since the twenty-first century,under the rapid development of computer hardware level,machine learning and deep learning have ushered in another spring.At the same time,natural language processing technology has also attracted more attention,and it is also used in many fields,such as language translation,information retrieval,automatic question answering,etc.,and all have good performance.This thesis attempts to apply natural language processing to the field of Chinese ancient poetry analysis,combined with deep learning algorithms to analyze and process the content of ancient poetry,and identify the artistic conception of poetry,including the theme and emotion of poetry.The research method of this thesis is to first process the open source data set,clean and supplement the original poetry data in the data set,and finally use the data set for the training of word vector model and algorithm model.Before identifying the artistic conception of ancient poems,first select ancient poetry word segmentation tools,compare various word segmentation tools such as THULAC,Jieba,Jiayan,etc.,and compare their performance in ancient poetry word segmentation.Combined with the word segmentation results,Word2 Vec is used to train the word vector model.In the topic classification of ancient poems,the problem is transformed into a text classification problem.This thesis compares a variety of text classification algorithms and summarizes their accuracy in the topic classification of ancient poems and words.Using Bi LSTM model based on Attention mechanism to analyze the sentiment of ancient poems is analyzed.In addition,this thesis also uses a graph database to build a knowledge graph of ancient poems,which connects the information of ancient poems and can also be used for information retrieval.Finally,the model is integrated into the ancient poetry analysis system,which allows users to input poems and get the results of artistic conception analysis of the poems.At the same time,other poems can be retrieved according to relevant information.