Automatic Recognition of Tune Names of Song Ci-Poetry

摘要

By introducing computer technology into the research of Song Ci-Poetry, we propose an automatic identification model for tune names of Song Ci-poetry in this paper. Firstly, the texts in Complete Collection of Song Ci Poetry involving 73 tune names of Ci are selected as experimental samples. The number of Chinese characters, the number of sentences, the number of clauses with different character number, information of Level and Oblique Tones and author information are made up into the feature system of Ci-Poetry. And then they are represented in vectors with Vector Space Model (VSM). Based on Support Vector Machine, Naive Bayes and Random Forest algorithms, three classifiers are used to recognize tune names of Ci. The model performs very well in recognition experiments, which provides a reference for the reorganization and retrieval work of Song Ci-Poetry.