Estimation of the Kansei Information obtained from Musical Scores via Machine Learning Algorithms : - Classification of Tempo into Two Classes Using Only Information Available in Musical Scores -
Estimation of the Kansei Information obtained from Musical Scores via Machine Learning Algorithms : - Classification of Tempo into Two Classes Using Only Information Available in Musical Scores -
复制标题
DOI：
10.1109/icawst.2019.8923480
复制
发表时间：
2019-10
期刊：
影响因子：
--
通讯作者：
Satoshi Kawamura;Zhongda Liu;H. Yoshida
中科院分区：
文献类型：
--
作者：
Satoshi Kawamura;Zhongda Liu;H. Yoshida
这项研究调查了机器学习算法是否可以用于仅根据乐谱上的音符序列将克里思准确地分为两类。在此，通过查看乐谱手动估计的克里思经由感性（情感）信息处理来模拟。克里思阈值被设置为m = 120。结果表明，即使在成功的学习，算法表现出较低的识别率，而分类慢克里思类的评价数据和一些数据被错误地识别。相比之下，当从评估数据中分类快克里思类别时，这些算法显示出较高的识别率。这些算法在数据中没有显示出任何识别错误。
This study investigates whether machine learning algorithms can be used to accurately classify tempo into two classes based only on the musical note sequence written on musical scores. Herein, the tempo that is manually estimated by looking at the score is simulated via Kansei (emotional) information processing. The tempo threshold was set at ♩ = 120. Results showed that even after successful learning, the algorithms showed low recognition rates while classifying slow tempo class from the evaluation data and some data were erroneously recognized. In contrast, the algorithms showed high recognition rates when classifying fast tempo class from the evaluation data. The algorithms did not show any recognition error in the data.