Piano music score difficulty identification method based on lifting decision tree

Abstract

The invention belongs to the field of machine learning, higher accuracy and stability of piano music score difficulty level identification are obtained, reliable piano difficulty information is provided for piano teaching and student learning, and the user experience of a music score website is improved. The technical scheme adopted by the invention is as follows. According to the piano music score difficulty identification method based on the lifting decision tree, a learning algorithm xgboost model of a multi-classification lifting decision tree based on grid search is established, accuracydetection and optimization are performed on the established model by using a test set, and piano music score difficulty is classified by using the optimized model; wherein the decision tree is used asa primary function, the XGBoost model is composed of a plurality of decision trees, the later decision tree fits the previous residual error, and the finally obtained prediction value is the sum of test results of all decision trees. The method is mainly applied to piano music score difficulty automatic identification occasions.

Description

technical field [0001] The invention belongs to the field of machine learning, and in particular relates to an XGBoost (learning algorithm based on boosting decision tree) piano score difficulty recognition method. Background technique [0002] Recognition of the difficulty level of piano score refers to the use of an algorithm to automatically identify the difficulty level of a certain piano score and give users a reference. Huge amounts of piano sheet music are created every day, and there are already tons of piano sheet music in the history of music. However, how to find the score matching the learner's learning level from the huge piano score data is a big challenge. For professional piano learners, there are generally fixed advanced teaching materials, but it is not conducive to personalized learning to stimulate the enthusiasm and interest of learners. It is necessary to measure the difficulty level of the massive music scores on the Internet. For music amateurs, cho...

AI Technical Summary

Problems solved by technology

Clustering algorithm is an unsupervised classification algorithm. Although it can make full use of the natural distribution relationship between features and difficulty levels, it cannot use existing difficulty level labels as prior knowledge to help classification

Images