Research On Emotional Characteristics And Evolution Of Tang And Song Poetry Based On Text Mining

This paper,with text mining and text machine learning,did many text machine learning to two corpora respectively called All Tang Poetry and All Songci.The theme is about emotional characteristics and their evolution,the former mainly includes forms of literature,tone and part of speech,the latter has time,school and landscape.Moreover,the thesis explained reasonably for some research results from history and modern science.As for Songci,each form has totally different titles,but each title has many works.Moreover,the quantity of character differentiate for each title,but some Songci have different character quantity compared to their titles,the change is about one to three characters.According to the tone,the fourth tone is used the most,the least is the third one,the first and the second tone almost has the same quantity in Tang poetry and Songci.When speaking the fourth tone,mouth is blowing out,which is a expressing-sentiment behavior.As for part of speech,there are forty four kinds of them,only the half is common,noun and verb are in the top two.The first period in Tang dynasty is different from another three periods,the first period uniquely has the poetry about the life in palace,there is lots of sadness in another three periods of Tang dynasty.The South Song is more euphemistic than The North Song.About school,the pastoral landscape poetry,full of living sense,shows the landscape.The frontier school,in the prosperous time of Tang dynasty,is about the frontier scene,and it expresses patriotism.The New Yuefu school is smooth in expression,and the content is easy to be understood.About the historical poetry,there are lots of feature words of time,especially about the present moment,the historical poetry shows emotion of present moment through history.The Han-Meng school is grander than other four schools,and this school is artistically creative.Songci’s Haofang school is more powerful,and Wanyue school is subtler.About landscape,the Tang poetry shows delight towards spring,however,Songci is sentimental about spring.Both Tang poetry and Songci express sadness for autumn,but the Songci is subtler.Tang poetry and Songci has less works about summer and winter than spring and autumn.The main innovations,in this thesis,are as following.From the tone and the part of speech in Tang poetry and Songci and through labeling tone and part of speech in Tang poetry and Songci,this paper showed a mode in emotion expression by semantically contextual connection,which is a new trying.With word2 vec generation algorithm,this paper explored the relation among emotional characteristics both in Tang poetry and Songci,which is also a new research mode.As for time,school and landscape in Tang poetry and Songci,with text segmentation and extraction of emotional words,this thesis found that the emotion change over time,school and landscape,which obviously made a new mode of studying ancient text like Tang poetry and Songci.