Modeling Emotional Patterns and Singing Styles of Ancient Poetry in Vocal Art Songs Based on Text Mining

Ancient poems and art songs are the treasures of Chinese civilization, and it is of great significance to sing and promote ancient poems and art songs for cultural inheritance and artistic innovation. This paper analyzes the word frequency of ancient poems from the Tang and Song dynasties and the emotional expression of the songs. Through the use of KH coder software for word frequency text mining of poetry word frequency, 75 high-frequency words are extracted, and the central vocabulary for generating song elements is calculated, resulting in a central score of 0.051 for “sun and moon”, which is higher than other words. At the same time, the K-Means clustering method is used to cluster word frequencies by word frequency type. There are 8 types of clusters, and 4 representative types are selected for analysis. The results show that the word frequency of ancient poetic songs selects more words of scenery, weather and landscape, indicating that poetic songs are good at borrowing things to express emotions, and use the scenery and other entities to accentuate the emotions of poetic songs. The Tang and Song dynasties’ poems are examined for genre and tone, and emotional patterns in poetic songs are examined in two ways, with the conclusion that emotions are mainly expressed psychologically. Finally, the singing style model is constructed according to four styles: chant, ethnic, American, and pop.